rollingscan very clean

likelyhood -> find similar
REFRESHCLAUD.MD 20260126

.idea/deploymentTargetSelector.xml

@@ -4,10 +4,10 @@
     <selectionStates>
       <SelectionState runConfigName="app">
         <option name="selectionMode" value="DROPDOWN" />
-        <DropdownSelection timestamp="2026-01-25T20:45:06.118763497Z">
+        <DropdownSelection timestamp="2026-01-27T00:21:15.014661014Z">
           <Target type="DEFAULT_BOOT">
             <handle>
-              <DeviceId pluginId="LocalEmulator" identifier="path=/home/genki/.android/avd/Medium_Phone.avd" />
+              <DeviceId pluginId="PhysicalDevice" identifier="serial=R3CX106YYCB" />
             </handle>
           </Target>
         </DropdownSelection>

app/build.gradle.kts

@@ -48,6 +48,9 @@ dependencies {
     implementation(libs.androidx.lifecycle.viewmodel.compose)
     implementation(libs.androidx.activity.compose)
 
+    // DataStore Preferences
+    implementation("androidx.datastore:datastore-preferences:1.1.1")
+
     // Compose
     implementation(platform(libs.androidx.compose.bom))
     implementation(libs.androidx.compose.ui)

app/src/main/java/com/placeholder/sherpai2/data/local/AppDatabase.kt

@@ -10,6 +10,10 @@ import com.placeholder.sherpai2.data.local.entity.*
 /**
  * AppDatabase - Complete database for SherpAI2
  *
+ * VERSION 12 - Distribution-based rejection stats
+ * - Added similarityStdDev, similarityMin to FaceModelEntity
+ * - Enables self-calibrating threshold for face matching
+ *
  * VERSION 10 - User Feedback Loop
  * - Added UserFeedbackEntity for storing user corrections
  * - Enables cluster refinement before training
@@ -44,14 +48,15 @@ import com.placeholder.sherpai2.data.local.entity.*
         PhotoFaceTagEntity::class,
         PersonAgeTagEntity::class,
         FaceCacheEntity::class,
-        UserFeedbackEntity::class,      // NEW: User corrections
+        UserFeedbackEntity::class,
+        PersonStatisticsEntity::class,   // Pre-computed person stats
 
         // ===== COLLECTIONS =====
         CollectionEntity::class,
         CollectionImageEntity::class,
         CollectionFilterEntity::class
     ],
-    version = 10,  // INCREMENTED for user feedback
+    version = 12,  // INCREMENTED for distribution-based rejection stats
     exportSchema = false
 )
 abstract class AppDatabase : RoomDatabase() {
@@ -70,7 +75,8 @@ abstract class AppDatabase : RoomDatabase() {
     abstract fun photoFaceTagDao(): PhotoFaceTagDao
     abstract fun personAgeTagDao(): PersonAgeTagDao
     abstract fun faceCacheDao(): FaceCacheDao
-    abstract fun userFeedbackDao(): UserFeedbackDao  // NEW
+    abstract fun userFeedbackDao(): UserFeedbackDao
+    abstract fun personStatisticsDao(): PersonStatisticsDao
 
     // ===== COLLECTIONS DAO =====
     abstract fun collectionDao(): CollectionDao
@@ -242,13 +248,60 @@ val MIGRATION_9_10 = object : Migration(9, 10) {
     }
 }
 
+/**
+ * MIGRATION 10 → 11 (Person Statistics)
+ *
+ * Changes:
+ * 1. Create person_statistics table for pre-computed aggregates
+ */
+val MIGRATION_10_11 = object : Migration(10, 11) {
+    override fun migrate(database: SupportSQLiteDatabase) {
+
+        // Create person_statistics table
+        database.execSQL("""
+            CREATE TABLE IF NOT EXISTS person_statistics (
+                personId TEXT PRIMARY KEY NOT NULL,
+                photoCount INTEGER NOT NULL DEFAULT 0,
+                firstPhotoDate INTEGER NOT NULL DEFAULT 0,
+                lastPhotoDate INTEGER NOT NULL DEFAULT 0,
+                averageConfidence REAL NOT NULL DEFAULT 0,
+                agesWithPhotos TEXT,
+                updatedAt INTEGER NOT NULL DEFAULT 0,
+                FOREIGN KEY(personId) REFERENCES persons(id) ON DELETE CASCADE
+            )
+        """)
+
+        // Index for sorting by photo count (People Dashboard)
+        database.execSQL("CREATE INDEX IF NOT EXISTS index_person_statistics_photoCount ON person_statistics(photoCount)")
+    }
+}
+
+/**
+ * MIGRATION 11 → 12 (Distribution-based Rejection Stats)
+ *
+ * Changes:
+ * 1. Add similarityStdDev column to face_models (default 0.05)
+ * 2. Add similarityMin column to face_models (default 0.6)
+ *
+ * These fields enable self-calibrating thresholds during scanning.
+ * During training, we compute stats from training sample similarities
+ * and use (mean - 2*stdDev) as a floor for matching.
+ */
+val MIGRATION_11_12 = object : Migration(11, 12) {
+    override fun migrate(database: SupportSQLiteDatabase) {
+        // Add distribution stats columns with sensible defaults for existing models
+        database.execSQL("ALTER TABLE face_models ADD COLUMN similarityStdDev REAL NOT NULL DEFAULT 0.05")
+        database.execSQL("ALTER TABLE face_models ADD COLUMN similarityMin REAL NOT NULL DEFAULT 0.6")
+    }
+}
+
 /**
  * PRODUCTION MIGRATION NOTES:
  *
  * Before shipping to users, update DatabaseModule to use migrations:
  *
  * Room.databaseBuilder(context, AppDatabase::class.java, "sherpai.db")
- *     .addMigrations(MIGRATION_7_8, MIGRATION_8_9, MIGRATION_9_10)  // Add all migrations
+ *     .addMigrations(MIGRATION_7_8, MIGRATION_8_9, MIGRATION_9_10, MIGRATION_10_11, MIGRATION_11_12)  // Add all migrations
  *     // .fallbackToDestructiveMigration()  // Remove this
  *     .build()
  */

app/src/main/java/com/placeholder/sherpai2/data/local/dao/Facecachedao.kt

@@ -233,6 +233,33 @@ interface FaceCacheDao {
         limit: Int = 500
     ): List<FaceCacheEntity>
 
+    /**
+     * Get premium face CANDIDATES - same criteria but WITHOUT embedding requirement.
+     * Used to find faces that need embedding generation.
+     */
+    @Query("""
+        SELECT fc.* FROM face_cache fc
+        INNER JOIN images i ON fc.imageId = i.imageId
+        WHERE i.faceCount = 1
+          AND fc.faceAreaRatio >= :minAreaRatio
+          AND fc.isFrontal = 1
+          AND fc.qualityScore >= :minQuality
+          AND fc.embedding IS NULL
+        ORDER BY fc.qualityScore DESC, fc.faceAreaRatio DESC
+        LIMIT :limit
+    """)
+    suspend fun getPremiumFaceCandidatesNeedingEmbeddings(
+        minAreaRatio: Float = 0.10f,
+        minQuality: Float = 0.7f,
+        limit: Int = 500
+    ): List<FaceCacheEntity>
+
+    /**
+     * Update embedding for a face cache entry
+     */
+    @Query("UPDATE face_cache SET embedding = :embedding WHERE imageId = :imageId AND faceIndex = :faceIndex")
+    suspend fun updateEmbedding(imageId: String, faceIndex: Int, embedding: String)
+
     /**
      * Count of premium faces available
      */

app/src/main/java/com/placeholder/sherpai2/data/local/dao/ImageDao.kt

@@ -66,6 +66,9 @@ interface ImageDao {
     @Query("SELECT * FROM images WHERE imageId = :imageId")
     suspend fun getImageById(imageId: String): ImageEntity?
 
+    @Query("SELECT * FROM images WHERE imageUri = :uri LIMIT 1")
+    suspend fun getImageByUri(uri: String): ImageEntity?
+
     /**
      * Stream images ordered by capture time (newest first).
      *

app/src/main/java/com/placeholder/sherpai2/data/local/dao/Photofacetagdao.kt

@@ -83,9 +83,89 @@ interface PhotoFaceTagDao {
      */
     @Query("SELECT * FROM photo_face_tags ORDER BY detectedAt DESC LIMIT :limit")
     suspend fun getRecentlyDetectedFaces(limit: Int): List<PhotoFaceTagEntity>
+
+    // ===== CO-OCCURRENCE QUERIES =====
+
+    /**
+     * Find people who appear in photos together with a given person.
+     * Returns list of (otherFaceModelId, count) sorted by count descending.
+     * Use case: "Who appears most with Mom?" or "Show photos of Mom WITH Dad"
+     */
+    @Query("""
+        SELECT pft2.faceModelId as otherFaceModelId, COUNT(DISTINCT pft1.imageId) as coCount
+        FROM photo_face_tags pft1
+        INNER JOIN photo_face_tags pft2 ON pft1.imageId = pft2.imageId
+        WHERE pft1.faceModelId = :faceModelId
+        AND pft2.faceModelId != :faceModelId
+        GROUP BY pft2.faceModelId
+        ORDER BY coCount DESC
+    """)
+    suspend fun getCoOccurrences(faceModelId: String): List<PersonCoOccurrence>
+
+    /**
+     * Get images where BOTH people appear together.
+     */
+    @Query("""
+        SELECT DISTINCT pft1.imageId
+        FROM photo_face_tags pft1
+        INNER JOIN photo_face_tags pft2 ON pft1.imageId = pft2.imageId
+        WHERE pft1.faceModelId = :faceModelId1
+        AND pft2.faceModelId = :faceModelId2
+        ORDER BY pft1.detectedAt DESC
+    """)
+    suspend fun getImagesWithBothPeople(faceModelId1: String, faceModelId2: String): List<String>
+
+    /**
+     * Get images where person appears ALONE (no other trained faces).
+     */
+    @Query("""
+        SELECT imageId FROM photo_face_tags
+        WHERE faceModelId = :faceModelId
+        AND imageId NOT IN (
+            SELECT imageId FROM photo_face_tags
+            WHERE faceModelId != :faceModelId
+        )
+        ORDER BY detectedAt DESC
+    """)
+    suspend fun getImagesWithPersonAlone(faceModelId: String): List<String>
+
+    /**
+     * Get images where ALL specified people appear (N-way intersection).
+     * For "Intersection Search" moonshot feature.
+     */
+    @Query("""
+        SELECT imageId FROM photo_face_tags
+        WHERE faceModelId IN (:faceModelIds)
+        GROUP BY imageId
+        HAVING COUNT(DISTINCT faceModelId) = :requiredCount
+    """)
+    suspend fun getImagesWithAllPeople(faceModelIds: List<String>, requiredCount: Int): List<String>
+
+    /**
+     * Get images with at least N of the specified people (family portrait detection).
+     */
+    @Query("""
+        SELECT imageId, COUNT(DISTINCT faceModelId) as memberCount
+        FROM photo_face_tags
+        WHERE faceModelId IN (:faceModelIds)
+        GROUP BY imageId
+        HAVING memberCount >= :minMembers
+        ORDER BY memberCount DESC
+    """)
+    suspend fun getFamilyPortraits(faceModelIds: List<String>, minMembers: Int): List<FamilyPortraitResult>
 }
 
+data class FamilyPortraitResult(
+    val imageId: String,
+    val memberCount: Int
+)
+
 data class FaceModelPhotoCount(
     val faceModelId: String,
     val photoCount: Int
 )
+
+data class PersonCoOccurrence(
+    val otherFaceModelId: String,
+    val coCount: Int
+)

app/src/main/java/com/placeholder/sherpai2/data/local/entity/Facecacheentity.kt

@@ -99,6 +99,13 @@ data class FaceCacheEntity(
     companion object {
         const val CURRENT_CACHE_VERSION = 1
 
+        /**
+         * Convert FloatArray embedding to JSON string for storage
+         */
+        fun embeddingToJson(embedding: FloatArray): String {
+            return embedding.joinToString(",")
+        }
+
         /**
          * Create from ML Kit face detection result
          */

app/src/main/java/com/placeholder/sherpai2/data/local/entity/Facerecognitionentities.kt

@@ -143,6 +143,13 @@ data class FaceModelEntity(
     @ColumnInfo(name = "averageConfidence")
     val averageConfidence: Float,
 
+    // Distribution stats for self-calibrating rejection
+    @ColumnInfo(name = "similarityStdDev")
+    val similarityStdDev: Float = 0.05f,  // Default for backwards compat
+
+    @ColumnInfo(name = "similarityMin")
+    val similarityMin: Float = 0.6f,      // Default for backwards compat
+
     @ColumnInfo(name = "createdAt")
     val createdAt: Long,
 
@@ -157,26 +164,29 @@ data class FaceModelEntity(
 ) {
     companion object {
         /**
-         * Backwards compatible create() method
-         * Used by existing FaceRecognitionRepository code
+         * Create with distribution stats for self-calibrating rejection
          */
         fun create(
             personId: String,
             embeddingArray: FloatArray,
             trainingImageCount: Int,
-            averageConfidence: Float
+            averageConfidence: Float,
+            similarityStdDev: Float = 0.05f,
+            similarityMin: Float = 0.6f
         ): FaceModelEntity {
-            return createFromEmbedding(personId, embeddingArray, trainingImageCount, averageConfidence)
+            return createFromEmbedding(personId, embeddingArray, trainingImageCount, averageConfidence, similarityStdDev, similarityMin)
         }
 
         /**
-         * Create from single embedding (backwards compatible)
+         * Create from single embedding with distribution stats
          */
         fun createFromEmbedding(
             personId: String,
             embeddingArray: FloatArray,
             trainingImageCount: Int,
-            averageConfidence: Float
+            averageConfidence: Float,
+            similarityStdDev: Float = 0.05f,
+            similarityMin: Float = 0.6f
         ): FaceModelEntity {
             val now = System.currentTimeMillis()
             val centroid = TemporalCentroid(
@@ -194,6 +204,8 @@ data class FaceModelEntity(
                 centroidsJson = serializeCentroids(listOf(centroid)),
                 trainingImageCount = trainingImageCount,
                 averageConfidence = averageConfidence,
+                similarityStdDev = similarityStdDev,
+                similarityMin = similarityMin,
                 createdAt = now,
                 updatedAt = now,
                 lastUsed = null,

app/src/main/java/com/placeholder/sherpai2/data/repository/Facerecognitionrepository.kt

@@ -2,8 +2,10 @@ package com.placeholder.sherpai2.data.repository
 
 import android.content.Context
 import android.graphics.Bitmap
+import android.util.Log
 import com.placeholder.sherpai2.data.local.dao.FaceModelDao
 import com.placeholder.sherpai2.data.local.dao.ImageDao
+import com.placeholder.sherpai2.data.local.dao.PersonAgeTagDao
 import com.placeholder.sherpai2.data.local.dao.PersonDao
 import com.placeholder.sherpai2.data.local.dao.PhotoFaceTagDao
 import com.placeholder.sherpai2.data.local.entity.*
@@ -31,8 +33,12 @@ class FaceRecognitionRepository @Inject constructor(
     private val personDao: PersonDao,
     private val imageDao: ImageDao,
     private val faceModelDao: FaceModelDao,
-    private val photoFaceTagDao: PhotoFaceTagDao
+    private val photoFaceTagDao: PhotoFaceTagDao,
+    private val personAgeTagDao: PersonAgeTagDao
 ) {
+    companion object {
+        private const val TAG = "FaceRecognitionRepo"
+    }
 
     private val faceNetModel by lazy { FaceNetModel(context) }
 
@@ -93,11 +99,19 @@ class FaceRecognitionRepository @Inject constructor(
         }
         val avgConfidence = confidences.average().toFloat()
 
+        // Compute distribution stats for self-calibrating rejection
+        val stdDev = kotlin.math.sqrt(
+            confidences.map { (it - avgConfidence).toDouble().let { d -> d * d } }.average()
+        ).toFloat()
+        val minSimilarity = confidences.minOrNull() ?: 0f
+
         val faceModel = FaceModelEntity.create(
             personId = personId,
             embeddingArray = personEmbedding,
             trainingImageCount = validImages.size,
-            averageConfidence = avgConfidence
+            averageConfidence = avgConfidence,
+            similarityStdDev = stdDev,
+            similarityMin = minSimilarity
         )
 
         faceModelDao.insertFaceModel(faceModel)
@@ -181,12 +195,15 @@ class FaceRecognitionRepository @Inject constructor(
             var highestSimilarity = threshold
 
             for (faceModel in faceModels) {
-                val modelEmbedding = faceModel.getEmbeddingArray()
-                val similarity = faceNetModel.calculateSimilarity(faceEmbedding, modelEmbedding)
+                // Check ALL centroids for best match (critical for children with age centroids)
+                val centroids = faceModel.getCentroids()
+                val bestCentroidSimilarity = centroids.maxOfOrNull { centroid ->
+                    faceNetModel.calculateSimilarity(faceEmbedding, centroid.getEmbeddingArray())
+                } ?: 0f
 
-                if (similarity > highestSimilarity) {
-                    highestSimilarity = similarity
-                    bestMatch = Pair(faceModel.id, similarity)
+                if (bestCentroidSimilarity > highestSimilarity) {
+                    highestSimilarity = bestCentroidSimilarity
+                    bestMatch = Pair(faceModel.id, bestCentroidSimilarity)
                 }
             }
 
@@ -374,9 +391,49 @@ class FaceRecognitionRepository @Inject constructor(
             onProgress = onProgress
         )
 
+        // Generate age tags for children
+        if (person.isChild && person.dateOfBirth != null) {
+            generateAgeTagsForTraining(person, validImages)
+        }
+
         person.id
     }
 
+    /**
+     * Generate age tags from training images for a child
+     */
+    private suspend fun generateAgeTagsForTraining(
+        person: PersonEntity,
+        validImages: List<TrainingSanityChecker.ValidTrainingImage>
+    ) {
+        try {
+            val dob = person.dateOfBirth ?: return
+
+            val tags = validImages.mapNotNull { img ->
+                val imageEntity = imageDao.getImageByUri(img.uri.toString()) ?: return@mapNotNull null
+                val ageMs = imageEntity.capturedAt - dob
+                val ageYears = (ageMs / (365.25 * 24 * 60 * 60 * 1000)).toInt()
+
+                if (ageYears < 0 || ageYears > 25) return@mapNotNull null
+
+                PersonAgeTagEntity.create(
+                    personId = person.id,
+                    personName = person.name,
+                    imageId = imageEntity.imageId,
+                    ageAtCapture = ageYears,
+                    confidence = 1.0f
+                )
+            }
+
+            if (tags.isNotEmpty()) {
+                personAgeTagDao.insertTags(tags)
+                Log.d(TAG, "Created ${tags.size} age tags for ${person.name}")
+            }
+        } catch (e: Exception) {
+            Log.e(TAG, "Failed to generate age tags", e)
+        }
+    }
+
     /**
      * Get face model by ID
      */

app/src/main/java/com/placeholder/sherpai2/di/RepositoryModule.kt

@@ -61,14 +61,16 @@ abstract class RepositoryModule {
             personDao: PersonDao,
             imageDao: ImageDao,
             faceModelDao: FaceModelDao,
-            photoFaceTagDao: PhotoFaceTagDao
+            photoFaceTagDao: PhotoFaceTagDao,
+            personAgeTagDao: PersonAgeTagDao
         ): FaceRecognitionRepository {
             return FaceRecognitionRepository(
                 context = context,
                 personDao = personDao,
                 imageDao = imageDao,
                 faceModelDao = faceModelDao,
-                photoFaceTagDao = photoFaceTagDao
+                photoFaceTagDao = photoFaceTagDao,
+                personAgeTagDao = personAgeTagDao
             )
         }
 

app/src/main/java/com/placeholder/sherpai2/domain/clustering/Faceclusteringservice.kt

@@ -15,6 +15,7 @@ import com.placeholder.sherpai2.data.local.dao.ImageDao
 import com.placeholder.sherpai2.data.local.entity.FaceCacheEntity
 import com.placeholder.sherpai2.data.local.entity.ImageEntity
 import com.placeholder.sherpai2.ml.FaceNetModel
+import com.placeholder.sherpai2.ml.FaceNormalizer
 import com.placeholder.sherpai2.ui.discover.DiscoverySettings
 import dagger.hilt.android.qualifiers.ApplicationContext
 import kotlinx.coroutines.Dispatchers
@@ -344,14 +345,9 @@ class FaceClusteringService @Inject constructor(
                                 }
 
                                 try {
-                                    // Crop and generate embedding
-                                    val faceBitmap = Bitmap.createBitmap(
-                                        bitmap,
-                                        mlFace.boundingBox.left.coerceIn(0, bitmap.width - 1),
-                                        mlFace.boundingBox.top.coerceIn(0, bitmap.height - 1),
-                                        mlFace.boundingBox.width().coerceAtMost(bitmap.width - mlFace.boundingBox.left),
-                                        mlFace.boundingBox.height().coerceAtMost(bitmap.height - mlFace.boundingBox.top)
-                                    )
+                                    // Crop and normalize face
+                                    val faceBitmap = FaceNormalizer.cropAndNormalize(bitmap, mlFace)
+                                        ?: return@forEach
 
                                     val embedding = faceNetModel.generateEmbedding(faceBitmap)
                                     faceBitmap.recycle()
@@ -591,13 +587,8 @@ class FaceClusteringService @Inject constructor(
                                 if (!qualityCheck.isValid) return@mapNotNull null
 
                                 try {
-                                    val faceBitmap = Bitmap.createBitmap(
-                                        bitmap,
-                                        face.boundingBox.left.coerceIn(0, bitmap.width - 1),
-                                        face.boundingBox.top.coerceIn(0, bitmap.height - 1),
-                                        face.boundingBox.width().coerceAtMost(bitmap.width - face.boundingBox.left),
-                                        face.boundingBox.height().coerceAtMost(bitmap.height - face.boundingBox.top)
-                                    )
+                                    val faceBitmap = FaceNormalizer.cropAndNormalize(bitmap, face)
+                                        ?: return@mapNotNull null
 
                                     val embedding = faceNetModel.generateEmbedding(faceBitmap)
                                     faceBitmap.recycle()

app/src/main/java/com/placeholder/sherpai2/domain/clustering/Facequalityfilter.kt

@@ -29,6 +29,64 @@ import kotlin.math.sqrt
  */
 object FaceQualityFilter {
 
+    /**
+     * Age group estimation for filtering (child vs adult detection)
+     */
+    enum class AgeGroup { CHILD, ADULT, UNCERTAIN }
+
+    /**
+     * Estimate whether a face belongs to a child or adult based on facial proportions.
+     *
+     * Uses two heuristics:
+     * 1. Eye position ratio - Children have larger foreheads, so eyes are lower (~45% from top)
+     *    Adults have eyes at ~35% from top
+     * 2. Face roundness (width/height ratio) - Children: ~0.85-1.0, Adults: ~0.7-0.85
+     *
+     * @return AgeGroup.CHILD, AgeGroup.ADULT, or AgeGroup.UNCERTAIN
+     */
+    fun estimateAgeGroup(face: Face, imageWidth: Int, imageHeight: Int): AgeGroup {
+        val leftEye = face.getLandmark(FaceLandmark.LEFT_EYE)
+        val rightEye = face.getLandmark(FaceLandmark.RIGHT_EYE)
+
+        if (leftEye == null || rightEye == null) {
+            return AgeGroup.UNCERTAIN
+        }
+
+        // Eye-to-face height ratio (where eyes sit relative to face top)
+        val faceHeight = face.boundingBox.height().toFloat()
+        val faceTop = face.boundingBox.top.toFloat()
+        val eyeY = (leftEye.position.y + rightEye.position.y) / 2
+        val eyePositionRatio = (eyeY - faceTop) / faceHeight
+
+        // Children: eyes at ~45% from top (larger forehead proportionally)
+        // Adults: eyes at ~35% from top
+        // Score: higher = more child-like
+
+        // Face roundness (width/height)
+        val faceWidth = face.boundingBox.width().toFloat()
+        val faceRatio = faceWidth / faceHeight
+        // Children: ratio ~0.85-1.0 (rounder faces)
+        // Adults: ratio ~0.7-0.85 (longer/narrower faces)
+
+        var childScore = 0
+
+        // Eye position scoring
+        if (eyePositionRatio > 0.45f) childScore += 2  // Strong child signal
+        else if (eyePositionRatio > 0.42f) childScore += 1  // Mild child signal
+        else if (eyePositionRatio < 0.35f) childScore -= 1  // Adult signal
+
+        // Face roundness scoring
+        if (faceRatio > 0.90f) childScore += 2  // Very round = child
+        else if (faceRatio > 0.82f) childScore += 1  // Somewhat round
+        else if (faceRatio < 0.75f) childScore -= 1  // Long face = adult
+
+        return when {
+            childScore >= 3 -> AgeGroup.CHILD
+            childScore <= 0 -> AgeGroup.ADULT
+            else -> AgeGroup.UNCERTAIN
+        }
+    }
+
     /**
      * Validate face for Discovery/Clustering
      *

app/src/main/java/com/placeholder/sherpai2/domain/usecase/Populatefacedetectioncacheusecase.kt

@@ -75,7 +75,21 @@ class PopulateFaceDetectionCacheUseCase @Inject constructor(
         )
 
         try {
-            val imagesToScan = imageDao.getImagesNeedingFaceDetection()
+            // Get images that need face detection (hasFaces IS NULL)
+            var imagesToScan = imageDao.getImagesNeedingFaceDetection()
+
+            // CRITICAL FIX: Also check for images marked as having faces but no FaceCacheEntity
+            if (imagesToScan.isEmpty()) {
+                val faceStats = faceCacheDao.getCacheStats()
+                if (faceStats.totalFaces == 0) {
+                    // FaceCacheEntity is empty - rescan images that have faces
+                    val imagesWithFaces = imageDao.getImagesWithFaces()
+                    if (imagesWithFaces.isNotEmpty()) {
+                        Log.w(TAG, "FaceCacheEntity empty but ${imagesWithFaces.size} images have faces - rescanning")
+                        imagesToScan = imagesWithFaces
+                    }
+                }
+            }
 
             if (imagesToScan.isEmpty()) {
                 Log.d(TAG, "No images need scanning")
@@ -184,7 +198,7 @@ class PopulateFaceDetectionCacheUseCase @Inject constructor(
                 imageUri = image.imageUri
             )
 
-            // Create FaceCacheEntity entries for each face
+            // Create FaceCacheEntity entries for each face (NO embeddings - generated on demand)
             val faceCacheEntries = faces.mapIndexed { index, face ->
                 createFaceCacheEntry(
                     imageId = image.imageId,
@@ -205,7 +219,8 @@ class PopulateFaceDetectionCacheUseCase @Inject constructor(
     /**
      * Create FaceCacheEntity from ML Kit Face
      *
-     * Uses FaceCacheEntity.create() which calculates quality metrics automatically
+     * Uses FaceCacheEntity.create() which calculates quality metrics automatically.
+     * Embeddings are NOT generated here - they're generated on-demand in Training/Discovery.
      */
     private fun createFaceCacheEntry(
         imageId: String,
@@ -225,7 +240,7 @@ class PopulateFaceDetectionCacheUseCase @Inject constructor(
             imageHeight = imageHeight,
             confidence = 0.9f,  // High confidence from accurate detector
             isFrontal = isFrontal,
-            embedding = null  // Will be generated later during Discovery
+            embedding = null  // Generated on-demand in Training/Discovery
         )
     }
 
@@ -312,13 +327,27 @@ class PopulateFaceDetectionCacheUseCase @Inject constructor(
         val imageStats = imageDao.getFaceCacheStats()
         val faceStats = faceCacheDao.getCacheStats()
 
+        // CRITICAL FIX: If ImageEntity says "scanned" but FaceCacheEntity is empty,
+        // we need to re-scan. This happens after DB migration clears face_cache table.
+        val imagesWithFaces = imageStats?.imagesWithFaces ?: 0
+        val facesCached = faceStats.totalFaces
+
+        // If we have images marked as having faces but no FaceCacheEntity entries,
+        // those images need re-scanning
+        val needsRescan = if (imagesWithFaces > 0 && facesCached == 0) {
+            Log.w(TAG, "⚠️ FaceCacheEntity is empty but $imagesWithFaces images marked as having faces - forcing rescan")
+            imagesWithFaces
+        } else {
+            imageStats?.needsScanning ?: 0
+        }
+
         CacheStats(
             totalImages = imageStats?.totalImages ?: 0,
             imagesWithFaceCache = imageStats?.imagesWithFaceCache ?: 0,
-            imagesWithFaces = imageStats?.imagesWithFaces ?: 0,
+            imagesWithFaces = imagesWithFaces,
             imagesWithoutFaces = imageStats?.imagesWithoutFaces ?: 0,
-            needsScanning = imageStats?.needsScanning ?: 0,
-            totalFacesCached = faceStats.totalFaces,
+            needsScanning = needsRescan,
+            totalFacesCached = facesCached,
             facesWithEmbeddings = faceStats.withEmbeddings,
             averageQuality = faceStats.avgQuality
         )

app/src/main/java/com/placeholder/sherpai2/ui/imagedetail/ImageDetailScreen.kt

@@ -20,6 +20,7 @@ import androidx.lifecycle.compose.collectAsStateWithLifecycle
 import androidx.navigation.NavController
 import coil.compose.AsyncImage
 import com.placeholder.sherpai2.data.local.entity.TagEntity
+import com.placeholder.sherpai2.ui.imagedetail.viewmodel.FaceTagInfo
 import com.placeholder.sherpai2.ui.imagedetail.viewmodel.ImageDetailViewModel
 import net.engawapg.lib.zoomable.rememberZoomState
 import net.engawapg.lib.zoomable.zoomable
@@ -51,8 +52,12 @@ fun ImageDetailScreen(
     }
 
     val tags by viewModel.tags.collectAsStateWithLifecycle()
+    val faceTags by viewModel.faceTags.collectAsStateWithLifecycle()
     var showTags by remember { mutableStateOf(false) }
 
+    // Total tag count for badge
+    val totalTagCount = tags.size + faceTags.size
+
     // Navigation state
     val currentIndex = if (allImageUris.isNotEmpty()) allImageUris.indexOf(imageUri) else -1
     val hasNavigation = allImageUris.isNotEmpty() && currentIndex >= 0
@@ -84,27 +89,35 @@ fun ImageDetailScreen(
                             horizontalArrangement = Arrangement.spacedBy(4.dp),
                             verticalAlignment = Alignment.CenterVertically
                         ) {
-                            if (tags.isNotEmpty()) {
+                            if (totalTagCount > 0) {
                                 Badge(
                                     containerColor = if (showTags)
                                         MaterialTheme.colorScheme.primary
+                                    else if (faceTags.isNotEmpty())
+                                        MaterialTheme.colorScheme.tertiary
                                     else
                                         MaterialTheme.colorScheme.surfaceVariant
                                 ) {
                                     Text(
-                                        tags.size.toString(),
+                                        totalTagCount.toString(),
                                         color = if (showTags)
                                             MaterialTheme.colorScheme.onPrimary
+                                        else if (faceTags.isNotEmpty())
+                                            MaterialTheme.colorScheme.onTertiary
                                         else
                                             MaterialTheme.colorScheme.onSurfaceVariant
                                     )
                                 }
                             }
                             Icon(
-                                if (showTags) Icons.Default.Label else Icons.Default.LocalOffer,
+                                if (faceTags.isNotEmpty()) Icons.Default.Face
+                                else if (showTags) Icons.Default.Label
+                                else Icons.Default.LocalOffer,
                                 "Show Tags",
                                 tint = if (showTags)
                                     MaterialTheme.colorScheme.primary
+                                else if (faceTags.isNotEmpty())
+                                    MaterialTheme.colorScheme.tertiary
                                 else
                                     MaterialTheme.colorScheme.onSurfaceVariant
                             )
@@ -189,6 +202,30 @@ fun ImageDetailScreen(
                         contentPadding = PaddingValues(16.dp),
                         verticalArrangement = Arrangement.spacedBy(8.dp)
                     ) {
+                        // Face Tags Section (People in Photo)
+                        if (faceTags.isNotEmpty()) {
+                            item {
+                                Text(
+                                    "People (${faceTags.size})",
+                                    style = MaterialTheme.typography.titleMedium,
+                                    fontWeight = FontWeight.Bold,
+                                    color = MaterialTheme.colorScheme.tertiary
+                                )
+                            }
+
+                            items(faceTags, key = { it.tagId }) { faceTag ->
+                                FaceTagCard(
+                                    faceTag = faceTag,
+                                    onRemove = { viewModel.removeFaceTag(faceTag) }
+                                )
+                            }
+
+                            item {
+                                Spacer(modifier = Modifier.height(8.dp))
+                            }
+                        }
+
+                        // Regular Tags Section
                         item {
                             Text(
                                 "Tags (${tags.size})",
@@ -197,7 +234,7 @@ fun ImageDetailScreen(
                             )
                         }
 
-                        if (tags.isEmpty()) {
+                        if (tags.isEmpty() && faceTags.isEmpty()) {
                             item {
                                 Text(
                                     "No tags yet",
@@ -205,6 +242,14 @@ fun ImageDetailScreen(
                                     color = MaterialTheme.colorScheme.onSurfaceVariant
                                 )
                             }
+                        } else if (tags.isEmpty()) {
+                            item {
+                                Text(
+                                    "No other tags",
+                                    style = MaterialTheme.typography.bodySmall,
+                                    color = MaterialTheme.colorScheme.onSurfaceVariant
+                                )
+                            }
                         }
 
                         items(tags, key = { it.tagId }) { tag ->
@@ -220,6 +265,83 @@ fun ImageDetailScreen(
     }
 }
 
+@Composable
+private fun FaceTagCard(
+    faceTag: FaceTagInfo,
+    onRemove: () -> Unit
+) {
+    Card(
+        modifier = Modifier.fillMaxWidth(),
+        colors = CardDefaults.cardColors(
+            containerColor = MaterialTheme.colorScheme.tertiaryContainer
+        ),
+        shape = RoundedCornerShape(8.dp)
+    ) {
+        Row(
+            modifier = Modifier
+                .fillMaxWidth()
+                .padding(12.dp),
+            horizontalArrangement = Arrangement.SpaceBetween,
+            verticalAlignment = Alignment.CenterVertically
+        ) {
+            Column(modifier = Modifier.weight(1f)) {
+                Row(
+                    horizontalArrangement = Arrangement.spacedBy(8.dp),
+                    verticalAlignment = Alignment.CenterVertically
+                ) {
+                    Icon(
+                        imageVector = Icons.Default.Face,
+                        contentDescription = null,
+                        modifier = Modifier.size(20.dp),
+                        tint = MaterialTheme.colorScheme.tertiary
+                    )
+                    Text(
+                        text = faceTag.personName,
+                        style = MaterialTheme.typography.bodyLarge,
+                        fontWeight = FontWeight.SemiBold
+                    )
+                }
+
+                Row(
+                    horizontalArrangement = Arrangement.spacedBy(4.dp),
+                    verticalAlignment = Alignment.CenterVertically
+                ) {
+                    Text(
+                        text = "Face Recognition",
+                        style = MaterialTheme.typography.labelSmall,
+                        color = MaterialTheme.colorScheme.onSurfaceVariant
+                    )
+                    Text(
+                        text = "•",
+                        style = MaterialTheme.typography.labelSmall,
+                        color = MaterialTheme.colorScheme.onSurfaceVariant
+                    )
+                    Text(
+                        text = "${(faceTag.confidence * 100).toInt()}% confidence",
+                        style = MaterialTheme.typography.labelSmall,
+                        color = if (faceTag.confidence >= 0.7f)
+                            MaterialTheme.colorScheme.primary
+                        else if (faceTag.confidence >= 0.5f)
+                            MaterialTheme.colorScheme.secondary
+                        else
+                            MaterialTheme.colorScheme.error
+                    )
+                }
+            }
+
+            // Remove button
+            IconButton(
+                onClick = onRemove,
+                colors = IconButtonDefaults.iconButtonColors(
+                    contentColor = MaterialTheme.colorScheme.error
+                )
+            ) {
+                Icon(Icons.Default.Delete, "Remove face tag")
+            }
+        }
+    }
+}
+
 @Composable
 private fun TagCard(
     tag: TagEntity,

app/src/main/java/com/placeholder/sherpai2/ui/imagedetail/viewmodel/ImageDetailViewModel.kt

@@ -2,6 +2,10 @@ package com.placeholder.sherpai2.ui.imagedetail.viewmodel
 
 import androidx.lifecycle.ViewModel
 import androidx.lifecycle.viewModelScope
+import com.placeholder.sherpai2.data.local.dao.FaceModelDao
+import com.placeholder.sherpai2.data.local.dao.ImageDao
+import com.placeholder.sherpai2.data.local.dao.PersonDao
+import com.placeholder.sherpai2.data.local.dao.PhotoFaceTagDao
 import com.placeholder.sherpai2.data.local.entity.TagEntity
 import com.placeholder.sherpai2.domain.repository.TaggingRepository
 import dagger.hilt.android.lifecycle.HiltViewModel
@@ -10,17 +14,33 @@ import kotlinx.coroutines.flow.*
 import kotlinx.coroutines.launch
 import javax.inject.Inject
 
+/**
+ * Represents a person tagged in this photo via face recognition
+ */
+data class FaceTagInfo(
+    val personId: String,
+    val personName: String,
+    val confidence: Float,
+    val faceModelId: String,
+    val tagId: String
+)
+
 /**
  * ImageDetailViewModel
  *
  * Owns:
  *  - Image context
  *  - Tag write operations
+ *  - Face tag display (people recognized in photo)
  */
 @HiltViewModel
 @OptIn(ExperimentalCoroutinesApi::class)
 class ImageDetailViewModel @Inject constructor(
-    private val tagRepository: TaggingRepository
+    private val tagRepository: TaggingRepository,
+    private val imageDao: ImageDao,
+    private val photoFaceTagDao: PhotoFaceTagDao,
+    private val faceModelDao: FaceModelDao,
+    private val personDao: PersonDao
 ) : ViewModel() {
 
     private val imageUri = MutableStateFlow<String?>(null)
@@ -37,8 +57,43 @@ class ImageDetailViewModel @Inject constructor(
                 initialValue = emptyList()
             )
 
+    // Face tags (people recognized in this photo)
+    private val _faceTags = MutableStateFlow<List<FaceTagInfo>>(emptyList())
+    val faceTags: StateFlow<List<FaceTagInfo>> = _faceTags.asStateFlow()
+
     fun loadImage(uri: String) {
         imageUri.value = uri
+        loadFaceTags(uri)
+    }
+
+    private fun loadFaceTags(uri: String) {
+        viewModelScope.launch {
+            try {
+                // Get imageId from URI
+                val image = imageDao.getImageByUri(uri) ?: return@launch
+
+                // Get face tags for this image
+                val faceTags = photoFaceTagDao.getTagsForImage(image.imageId)
+
+                // Resolve to person names
+                val faceTagInfos = faceTags.mapNotNull { tag ->
+                    val faceModel = faceModelDao.getFaceModelById(tag.faceModelId) ?: return@mapNotNull null
+                    val person = personDao.getPersonById(faceModel.personId) ?: return@mapNotNull null
+
+                    FaceTagInfo(
+                        personId = person.id,
+                        personName = person.name,
+                        confidence = tag.confidence,
+                        faceModelId = tag.faceModelId,
+                        tagId = tag.id
+                    )
+                }
+
+                _faceTags.value = faceTagInfos.sortedByDescending { it.confidence }
+            } catch (e: Exception) {
+                _faceTags.value = emptyList()
+            }
+        }
     }
 
     fun addTag(value: String) {
@@ -54,4 +109,15 @@ class ImageDetailViewModel @Inject constructor(
             tagRepository.removeTagFromImage(uri, tag.value)
         }
     }
+
+    /**
+     * Remove a face tag (person recognition)
+     */
+    fun removeFaceTag(faceTagInfo: FaceTagInfo) {
+        viewModelScope.launch {
+            photoFaceTagDao.deleteTagById(faceTagInfo.tagId)
+            // Reload face tags
+            imageUri.value?.let { loadFaceTags(it) }
+        }
+    }
 }

app/src/main/java/com/placeholder/sherpai2/ui/modelinventory/Personinventoryscreen.kt

@@ -95,6 +95,9 @@ fun PersonInventoryScreen(
                     },
                     onDelete = { personId ->
                         viewModel.deletePerson(personId)
+                    },
+                    onClearTags = { personId ->
+                        viewModel.clearTagsForPerson(personId)
                     }
                 )
             }
@@ -319,7 +322,8 @@ private fun PersonList(
     persons: List<PersonWithModelInfo>,
     onScan: (String) -> Unit,
     onView: (String) -> Unit,
-    onDelete: (String) -> Unit
+    onDelete: (String) -> Unit,
+    onClearTags: (String) -> Unit
 ) {
     LazyColumn(
         contentPadding = PaddingValues(vertical = 8.dp)
@@ -332,7 +336,8 @@ private fun PersonList(
                 person = person,
                 onScan = { onScan(person.person.id) },
                 onView = { onView(person.person.id) },
-                onDelete = { onDelete(person.person.id) }
+                onDelete = { onDelete(person.person.id) },
+                onClearTags = { onClearTags(person.person.id) }
             )
         }
     }
@@ -343,9 +348,34 @@ private fun PersonCard(
     person: PersonWithModelInfo,
     onScan: () -> Unit,
     onView: () -> Unit,
-    onDelete: () -> Unit
+    onDelete: () -> Unit,
+    onClearTags: () -> Unit
 ) {
     var showDeleteDialog by remember { mutableStateOf(false) }
+    var showClearDialog by remember { mutableStateOf(false) }
+
+    if (showClearDialog) {
+        AlertDialog(
+            onDismissRequest = { showClearDialog = false },
+            title = { Text("Clear tags for ${person.person.name}?") },
+            text = { Text("This will remove all ${person.taggedPhotoCount} photo tags but keep the face model. You can re-scan after clearing.") },
+            confirmButton = {
+                TextButton(
+                    onClick = {
+                        showClearDialog = false
+                        onClearTags()
+                    }
+                ) {
+                    Text("Clear Tags", color = MaterialTheme.colorScheme.error)
+                }
+            },
+            dismissButton = {
+                TextButton(onClick = { showClearDialog = false }) {
+                    Text("Cancel")
+                }
+            }
+        )
+    }
 
     if (showDeleteDialog) {
         AlertDialog(
@@ -413,6 +443,17 @@ private fun PersonCard(
                     )
                 }
 
+                // Clear tags button (if has tags)
+                if (person.taggedPhotoCount > 0) {
+                    IconButton(onClick = { showClearDialog = true }) {
+                        Icon(
+                            Icons.Default.ClearAll,
+                            contentDescription = "Clear Tags",
+                            tint = MaterialTheme.colorScheme.secondary
+                        )
+                    }
+                }
+
                 // Delete button
                 IconButton(onClick = { showDeleteDialog = true }) {
                     Icon(

app/src/main/java/com/placeholder/sherpai2/ui/modelinventory/Personinventoryviewmodel.kt

@@ -19,6 +19,7 @@ import com.placeholder.sherpai2.data.local.entity.PersonEntity
 import com.placeholder.sherpai2.data.local.entity.PhotoFaceTagEntity
 import com.placeholder.sherpai2.ml.FaceNetModel
 import com.placeholder.sherpai2.ml.ThresholdStrategy
+import com.placeholder.sherpai2.domain.clustering.FaceQualityFilter
 import dagger.hilt.android.lifecycle.HiltViewModel
 import dagger.hilt.android.qualifiers.ApplicationContext
 import kotlinx.coroutines.Dispatchers
@@ -105,6 +106,21 @@ class PersonInventoryViewModel @Inject constructor(
         }
     }
 
+    /**
+     * Clear all face tags for a person (keep model, allow rescan)
+     */
+    fun clearTagsForPerson(personId: String) {
+        viewModelScope.launch(Dispatchers.IO) {
+            try {
+                val faceModel = faceModelDao.getFaceModelByPersonId(personId)
+                if (faceModel != null) {
+                    photoFaceTagDao.deleteTagsForFaceModel(faceModel.id)
+                }
+                loadPersons()
+            } catch (e: Exception) {}
+        }
+    }
+
     fun scanForPerson(personId: String) {
         viewModelScope.launch(Dispatchers.IO) {
             try {
@@ -127,16 +143,40 @@ class PersonInventoryViewModel @Inject constructor(
 
                 val detectorOptions = FaceDetectorOptions.Builder()
                     .setPerformanceMode(FaceDetectorOptions.PERFORMANCE_MODE_ACCURATE)
-                    .setLandmarkMode(FaceDetectorOptions.LANDMARK_MODE_NONE)
+                    .setLandmarkMode(FaceDetectorOptions.LANDMARK_MODE_ALL)  // Needed for age estimation
                     .setClassificationMode(FaceDetectorOptions.CLASSIFICATION_MODE_NONE)
                     .setMinFaceSize(0.15f)
                     .build()
 
                 val detector = FaceDetection.getClient(detectorOptions)
-                val modelEmbedding = faceModel.getEmbeddingArray()
-                val faceNetModel = FaceNetModel(context)
+                // CRITICAL: Use ALL centroids for matching
+                val modelCentroids = faceModel.getCentroids().map { it.getEmbeddingArray() }
                 val trainingCount = faceModel.trainingImageCount
-                val baseThreshold = ThresholdStrategy.getLiberalThreshold(trainingCount)
+                android.util.Log.e("PersonScan", "=== CENTROIDS: ${modelCentroids.size}, trainingCount: $trainingCount ===")
+
+                if (modelCentroids.isEmpty()) {
+                    _scanningState.value = ScanningState.Error("No centroids found")
+                    return@launch
+                }
+
+                val faceNetModel = FaceNetModel(context)
+                // Production threshold - STRICT to avoid false positives
+                // Solo face photos: 0.62, Group photos: 0.68
+                val baseThreshold = 0.62f
+                val groupPhotoThreshold = 0.68f  // Higher bar for multi-face images
+
+                // Load ALL other models for "best match wins" comparison
+                val allModels = faceModelDao.getAllActiveFaceModels()
+                val otherModelCentroids = allModels
+                    .filter { it.id != faceModel.id }
+                    .map { model -> model.id to model.getCentroids().map { it.getEmbeddingArray() } }
+
+                // Distribution-based minimum threshold (self-calibrating)
+                val distributionMin = (faceModel.averageConfidence - 2 * faceModel.similarityStdDev)
+                    .coerceAtLeast(faceModel.similarityMin - 0.05f)
+                    .coerceAtLeast(0.50f)  // Never go below 0.50 absolute floor
+
+                android.util.Log.d("PersonScan", "Using threshold: solo=$baseThreshold, group=$groupPhotoThreshold, distributionMin=$distributionMin (avgConf=${faceModel.averageConfidence}, stdDev=${faceModel.similarityStdDev}), centroids: ${modelCentroids.size}, competing models: ${otherModelCentroids.size}, isChild=${person.isChild}")
 
                 val completed = AtomicInteger(0)
                 val facesFound = AtomicInteger(0)
@@ -148,7 +188,7 @@ class PersonInventoryViewModel @Inject constructor(
                     val jobs = untaggedImages.map { image ->
                         async {
                             semaphore.withPermit {
-                                processImage(image, detector, faceNetModel, modelEmbedding, trainingCount, baseThreshold, personId, faceModel.id, batchMatches, batchUpdateMutex, completed, facesFound, startTime, totalToScan, person.name)
+                                processImage(image, detector, faceNetModel, modelCentroids, otherModelCentroids, trainingCount, baseThreshold, groupPhotoThreshold, distributionMin, person.isChild, personId, faceModel.id, batchMatches, batchUpdateMutex, completed, facesFound, startTime, totalToScan, person.name)
                             }
                         }
                     }
@@ -175,7 +215,10 @@ class PersonInventoryViewModel @Inject constructor(
 
     private suspend fun processImage(
         image: ImageEntity, detector: com.google.mlkit.vision.face.FaceDetector, faceNetModel: FaceNetModel,
-        modelEmbedding: FloatArray, trainingCount: Int, baseThreshold: Float, personId: String, faceModelId: String,
+        modelCentroids: List<FloatArray>, otherModelCentroids: List<Pair<String, List<FloatArray>>>,
+        trainingCount: Int, baseThreshold: Float, groupPhotoThreshold: Float,
+        distributionMin: Float, isChildTarget: Boolean,
+        personId: String, faceModelId: String,
         batchMatches: MutableList<Triple<String, String, Float>>, batchUpdateMutex: Mutex,
         completed: AtomicInteger, facesFound: AtomicInteger, startTime: Long, totalToScan: Int, personName: String
     ) {
@@ -200,9 +243,13 @@ class PersonInventoryViewModel @Inject constructor(
             val scaleX = sizeOpts.outWidth.toFloat() / detectionBitmap.width
             val scaleY = sizeOpts.outHeight.toFloat() / detectionBitmap.height
 
-            val imageQuality = ThresholdStrategy.estimateImageQuality(sizeOpts.outWidth, sizeOpts.outHeight)
-            val detectionContext = ThresholdStrategy.estimateDetectionContext(faces.size)
-            val threshold = ThresholdStrategy.getOptimalThreshold(trainingCount, imageQuality, detectionContext).coerceAtMost(baseThreshold)
+            // CRITICAL: Use higher threshold for group photos (more likely false positives)
+            val isGroupPhoto = faces.size > 1
+            val effectiveThreshold = if (isGroupPhoto) groupPhotoThreshold else baseThreshold
+
+            // Track best match in this image (only tag ONE face per image)
+            var bestMatchSimilarity = 0f
+            var foundMatch = false
 
             for (face in faces) {
                 val scaledBounds = android.graphics.Rect(
@@ -212,14 +259,62 @@ class PersonInventoryViewModel @Inject constructor(
                     (face.boundingBox.bottom * scaleY).toInt()
                 )
 
-                val faceBitmap = loadFaceRegion(uri, scaledBounds) ?: continue
+                // Skip very small faces (less reliable)
+                val faceArea = scaledBounds.width() * scaledBounds.height()
+                val imageArea = sizeOpts.outWidth * sizeOpts.outHeight
+                val faceRatio = faceArea.toFloat() / imageArea
+                if (faceRatio < 0.02f) continue  // Face must be at least 2% of image
+
+                // SIGNAL 2: Age plausibility check (if target is a child)
+                if (isChildTarget) {
+                    val ageGroup = FaceQualityFilter.estimateAgeGroup(face, detectionBitmap.width, detectionBitmap.height)
+                    if (ageGroup == FaceQualityFilter.AgeGroup.ADULT) {
+                        continue  // Reject clearly adult faces when searching for a child
+                    }
+                }
+
+                // CRITICAL: Add padding to face crop (same as training)
+                val faceBitmap = loadFaceRegionWithPadding(uri, scaledBounds, sizeOpts.outWidth, sizeOpts.outHeight) ?: continue
                 val faceEmbedding = faceNetModel.generateEmbedding(faceBitmap)
-                val similarity = faceNetModel.calculateSimilarity(faceEmbedding, modelEmbedding)
                 faceBitmap.recycle()
 
-                if (similarity >= threshold) {
+                // Match against target person's centroids
+                val targetSimilarity = modelCentroids.maxOfOrNull { centroid ->
+                    faceNetModel.calculateSimilarity(faceEmbedding, centroid)
+                } ?: 0f
+
+                // SIGNAL 1: Distribution-based rejection
+                // If similarity is below (mean - 2*stdDev) or (min - 0.05), it's a statistical outlier
+                if (targetSimilarity < distributionMin) {
+                    continue  // Too far below training distribution
+                }
+
+                // SIGNAL 3: Basic threshold check
+                if (targetSimilarity < effectiveThreshold) {
+                    continue
+                }
+
+                // SIGNAL 4: "Best match wins" - check if any OTHER model scores higher
+                // This prevents tagging siblings/similar people incorrectly
+                val bestOtherSimilarity = otherModelCentroids.maxOfOrNull { (_, centroids) ->
+                    centroids.maxOfOrNull { centroid ->
+                        faceNetModel.calculateSimilarity(faceEmbedding, centroid)
+                    } ?: 0f
+                } ?: 0f
+
+                val isTargetBestMatch = targetSimilarity > bestOtherSimilarity
+
+                // All signals must pass
+                if (isTargetBestMatch && targetSimilarity > bestMatchSimilarity) {
+                    bestMatchSimilarity = targetSimilarity
+                    foundMatch = true
+                }
+            }
+
+            // Only add ONE tag per image (the best match)
+            if (foundMatch) {
                 batchUpdateMutex.withLock {
-                        batchMatches.add(Triple(personId, image.imageId, similarity))
+                    batchMatches.add(Triple(personId, image.imageId, bestMatchSimilarity))
                     facesFound.incrementAndGet()
                     if (batchMatches.size >= BATCH_DB_SIZE) {
                         saveBatchMatches(batchMatches.toList(), faceModelId)
@@ -227,7 +322,7 @@ class PersonInventoryViewModel @Inject constructor(
                     }
                 }
             }
-            }
+
             detectionBitmap.recycle()
         } catch (e: Exception) {
         } finally {
@@ -250,18 +345,32 @@ class PersonInventoryViewModel @Inject constructor(
         } catch (e: Exception) { null }
     }
 
-    private fun loadFaceRegion(uri: Uri, bounds: android.graphics.Rect): Bitmap? {
+    /**
+     * Load face region WITH 25% padding - CRITICAL for matching training conditions
+     */
+    private fun loadFaceRegionWithPadding(uri: Uri, bounds: android.graphics.Rect, imgWidth: Int, imgHeight: Int): Bitmap? {
         return try {
             val full = context.contentResolver.openInputStream(uri)?.use {
                 BitmapFactory.decodeStream(it, null, BitmapFactory.Options().apply { inPreferredConfig = Bitmap.Config.ARGB_8888 })
             } ?: return null
 
-            val safeLeft = bounds.left.coerceIn(0, full.width - 1)
-            val safeTop = bounds.top.coerceIn(0, full.height - 1)
-            val safeWidth = bounds.width().coerceAtMost(full.width - safeLeft)
-            val safeHeight = bounds.height().coerceAtMost(full.height - safeTop)
+            // Add 25% padding (same as training)
+            val padding = (kotlin.math.max(bounds.width(), bounds.height()) * 0.25f).toInt()
 
-            val cropped = Bitmap.createBitmap(full, safeLeft, safeTop, safeWidth, safeHeight)
+            val left = (bounds.left - padding).coerceAtLeast(0)
+            val top = (bounds.top - padding).coerceAtLeast(0)
+            val right = (bounds.right + padding).coerceAtMost(full.width)
+            val bottom = (bounds.bottom + padding).coerceAtMost(full.height)
+
+            val width = right - left
+            val height = bottom - top
+
+            if (width <= 0 || height <= 0) {
+                full.recycle()
+                return null
+            }
+
+            val cropped = Bitmap.createBitmap(full, left, top, width, height)
             full.recycle()
             cropped
         } catch (e: Exception) { null }

app/src/main/java/com/placeholder/sherpai2/ui/navigation/AppNavHost.kt

@@ -339,10 +339,7 @@ fun AppNavHost(
          * SETTINGS SCREEN
          */
         composable(AppRoutes.SETTINGS) {
-            DummyScreen(
-                title = "Settings",
-                subtitle = "App preferences and configuration"
-            )
+            com.placeholder.sherpai2.ui.settings.SettingsScreen()
         }
     }
 }

app/src/main/java/com/placeholder/sherpai2/ui/presentation/MainScreen.kt

@@ -78,6 +78,7 @@ fun MainScreen(
                                     AppRoutes.DISCOVER -> "Discover People"  // ✅ SHOWS NOW!
                                     AppRoutes.INVENTORY -> "People"
                                     AppRoutes.TRAIN -> "Train Model"
+                                    AppRoutes.ScanResultsScreen -> "Train New Person"
                                     AppRoutes.TAGS -> "Tags"
                                     AppRoutes.UTILITIES -> "Utilities"
                                     AppRoutes.SETTINGS -> "Settings"

app/src/main/java/com/placeholder/sherpai2/ui/rollingscan/Rollingscanscreen.kt

@@ -2,7 +2,9 @@ package com.placeholder.sherpai2.ui.rollingscan
 
 import android.net.Uri
 import androidx.compose.foundation.BorderStroke
+import androidx.compose.foundation.ExperimentalFoundationApi
 import androidx.compose.foundation.clickable
+import androidx.compose.foundation.combinedClickable
 import androidx.compose.foundation.layout.*
 import androidx.compose.foundation.lazy.grid.GridCells
 import androidx.compose.foundation.lazy.grid.GridItemSpan
@@ -37,7 +39,7 @@ import com.placeholder.sherpai2.domain.similarity.FaceSimilarityScorer
  * - Quick action buttons (Select Top N)
  * - Submit button with validation
  */
-@OptIn(ExperimentalMaterial3Api::class)
+@OptIn(ExperimentalMaterial3Api::class, ExperimentalFoundationApi::class)
 @Composable
 fun RollingScanScreen(
     seedImageIds: List<String>,
@@ -48,6 +50,7 @@ fun RollingScanScreen(
 ) {
     val uiState by viewModel.uiState.collectAsState()
     val selectedImageIds by viewModel.selectedImageIds.collectAsState()
+    val negativeImageIds by viewModel.negativeImageIds.collectAsState()
     val rankedPhotos by viewModel.rankedPhotos.collectAsState()
     val isScanning by viewModel.isScanning.collectAsState()
 
@@ -70,6 +73,7 @@ fun RollingScanScreen(
                 isReadyForTraining = viewModel.isReadyForTraining(),
                 validationMessage = viewModel.getValidationMessage(),
                 onSelectTopN = { count -> viewModel.selectTopN(count) },
+                onSelectAboveThreshold = { threshold -> viewModel.selectAllAboveThreshold(threshold) },
                 onSubmit = {
                     val uris = viewModel.getSelectedImageUris()
                     onSubmitForTraining(uris)
@@ -93,8 +97,10 @@ fun RollingScanScreen(
                 RollingScanPhotoGrid(
                     rankedPhotos = rankedPhotos,
                     selectedImageIds = selectedImageIds,
+                    negativeImageIds = negativeImageIds,
                     isScanning = isScanning,
                     onToggleSelection = { imageId -> viewModel.toggleSelection(imageId) },
+                    onToggleNegative = { imageId -> viewModel.toggleNegative(imageId) },
                     modifier = Modifier.padding(padding)
                 )
             }
@@ -159,19 +165,26 @@ private fun RollingScanTopBar(
 }
 
 // ═══════════════════════════════════════════════════════════
-// PHOTO GRID
+// PHOTO GRID - Similarity-based bucketing
 // ═══════════════════════════════════════════════════════════
 
+@OptIn(ExperimentalFoundationApi::class)
 @Composable
 private fun RollingScanPhotoGrid(
     rankedPhotos: List<FaceSimilarityScorer.ScoredPhoto>,
     selectedImageIds: Set<String>,
+    negativeImageIds: Set<String>,
     isScanning: Boolean,
     onToggleSelection: (String) -> Unit,
+    onToggleNegative: (String) -> Unit,
     modifier: Modifier = Modifier
 ) {
-    Column(modifier = modifier.fillMaxSize()) {
+    // Bucket by similarity score
+    val veryLikely = rankedPhotos.filter { it.finalScore >= 0.60f }
+    val probably = rankedPhotos.filter { it.finalScore in 0.45f..0.599f }
+    val maybe = rankedPhotos.filter { it.finalScore < 0.45f }
 
+    Column(modifier = modifier.fillMaxSize()) {
         // Scanning indicator
         if (isScanning) {
             LinearProgressIndicator(
@@ -180,69 +193,78 @@ private fun RollingScanPhotoGrid(
             )
         }
 
+        // Hint for negative marking
+        Text(
+            text = "Tap to select • Long-press to mark as NOT this person",
+            style = MaterialTheme.typography.bodySmall,
+            color = MaterialTheme.colorScheme.onSurfaceVariant,
+            modifier = Modifier.padding(horizontal = 12.dp, vertical = 4.dp)
+        )
+
         LazyVerticalGrid(
             columns = GridCells.Fixed(3),
             contentPadding = PaddingValues(8.dp),
             horizontalArrangement = Arrangement.spacedBy(8.dp),
             verticalArrangement = Arrangement.spacedBy(8.dp)
         ) {
-            // Section: Most Similar (top 10)
-            val topMatches = rankedPhotos.take(10)
-            if (topMatches.isNotEmpty()) {
+            // Section: Very Likely (>60%)
+            if (veryLikely.isNotEmpty()) {
                 item(span = { GridItemSpan(3) }) {
                     SectionHeader(
                         icon = Icons.Default.Whatshot,
-                        text = "🔥 Most Similar (${topMatches.size})",
-                        color = MaterialTheme.colorScheme.primary
+                        text = "🟢 Very Likely (${veryLikely.size})",
+                        color = Color(0xFF4CAF50)
                     )
                 }
-
-                items(topMatches, key = { it.imageId }) { photo ->
+                items(veryLikely, key = { it.imageId }) { photo ->
                     PhotoCard(
                         photo = photo,
                         isSelected = photo.imageId in selectedImageIds,
+                        isNegative = photo.imageId in negativeImageIds,
                         onToggle = { onToggleSelection(photo.imageId) },
+                        onLongPress = { onToggleNegative(photo.imageId) },
                         showSimilarityBadge = true
                     )
                 }
             }
 
-            // Section: Good Matches (11-30)
-            val goodMatches = rankedPhotos.drop(10).take(20)
-            if (goodMatches.isNotEmpty()) {
+            // Section: Probably (45-60%)
+            if (probably.isNotEmpty()) {
                 item(span = { GridItemSpan(3) }) {
                     SectionHeader(
                         icon = Icons.Default.CheckCircle,
-                        text = "📊 Good Matches (${goodMatches.size})",
-                        color = MaterialTheme.colorScheme.tertiary
+                        text = "🟡 Probably (${probably.size})",
+                        color = Color(0xFFFFC107)
                     )
                 }
-
-                items(goodMatches, key = { it.imageId }) { photo ->
+                items(probably, key = { it.imageId }) { photo ->
                     PhotoCard(
                         photo = photo,
                         isSelected = photo.imageId in selectedImageIds,
-                        onToggle = { onToggleSelection(photo.imageId) }
+                        isNegative = photo.imageId in negativeImageIds,
+                        onToggle = { onToggleSelection(photo.imageId) },
+                        onLongPress = { onToggleNegative(photo.imageId) },
+                        showSimilarityBadge = true
                     )
                 }
             }
 
-            // Section: Other Photos
-            val otherPhotos = rankedPhotos.drop(30)
-            if (otherPhotos.isNotEmpty()) {
+            // Section: Maybe (<45%)
+            if (maybe.isNotEmpty()) {
                 item(span = { GridItemSpan(3) }) {
                     SectionHeader(
                         icon = Icons.Default.Photo,
-                        text = "📷 Other Photos (${otherPhotos.size})",
-                        color = MaterialTheme.colorScheme.onSurfaceVariant
+                        text = "🟠 Maybe (${maybe.size})",
+                        color = Color(0xFFFF9800)
                     )
                 }
-
-                items(otherPhotos, key = { it.imageId }) { photo ->
+                items(maybe, key = { it.imageId }) { photo ->
                     PhotoCard(
                         photo = photo,
                         isSelected = photo.imageId in selectedImageIds,
-                        onToggle = { onToggleSelection(photo.imageId) }
+                        isNegative = photo.imageId in negativeImageIds,
+                        onToggle = { onToggleSelection(photo.imageId) },
+                        onLongPress = { onToggleNegative(photo.imageId) }
                     )
                 }
             }
@@ -258,24 +280,34 @@ private fun RollingScanPhotoGrid(
 }
 
 // ═══════════════════════════════════════════════════════════
-// PHOTO CARD
+// PHOTO CARD - with long-press for negative marking
 // ═══════════════════════════════════════════════════════════
 
+@OptIn(ExperimentalFoundationApi::class)
 @Composable
 private fun PhotoCard(
     photo: FaceSimilarityScorer.ScoredPhoto,
     isSelected: Boolean,
+    isNegative: Boolean = false,
     onToggle: () -> Unit,
+    onLongPress: () -> Unit = {},
     showSimilarityBadge: Boolean = false
 ) {
+    val borderColor = when {
+        isNegative -> Color(0xFFE53935)  // Red for negative
+        isSelected -> MaterialTheme.colorScheme.primary
+        else -> MaterialTheme.colorScheme.outline.copy(alpha = 0.3f)
+    }
+    val borderWidth = if (isSelected || isNegative) 3.dp else 1.dp
+
     Card(
         modifier = Modifier
             .aspectRatio(1f)
-            .clickable(onClick = onToggle),
-        border = if (isSelected)
-            BorderStroke(3.dp, MaterialTheme.colorScheme.primary)
-        else
-            BorderStroke(1.dp, MaterialTheme.colorScheme.outline.copy(alpha = 0.3f)),
+            .combinedClickable(
+                onClick = onToggle,
+                onLongClick = onLongPress
+            ),
+        border = BorderStroke(borderWidth, borderColor),
         elevation = CardDefaults.cardElevation(
             defaultElevation = if (isSelected) 4.dp else 1.dp
         )
@@ -289,22 +321,47 @@ private fun PhotoCard(
                 contentScale = ContentScale.Crop
             )
 
-            // Similarity badge (top-left) - Only for top matches
-            if (showSimilarityBadge) {
+            // Dim overlay for negatives
+            if (isNegative) {
+                Box(
+                    modifier = Modifier
+                        .fillMaxSize()
+                        .padding(0.dp),
+                    contentAlignment = Alignment.Center
+                ) {
+                    Surface(
+                        modifier = Modifier.fillMaxSize(),
+                        color = Color.Black.copy(alpha = 0.5f)
+                    ) {}
+                    Icon(
+                        Icons.Default.Close,
+                        contentDescription = "Not this person",
+                        tint = Color.White,
+                        modifier = Modifier.size(32.dp)
+                    )
+                }
+            }
+
+            // Similarity badge (top-left)
+            if (showSimilarityBadge && !isNegative) {
                 Surface(
                     modifier = Modifier
                         .align(Alignment.TopStart)
                         .padding(6.dp),
                     shape = RoundedCornerShape(8.dp),
-                    color = MaterialTheme.colorScheme.primary,
+                    color = when {
+                        photo.finalScore >= 0.60f -> Color(0xFF4CAF50)
+                        photo.finalScore >= 0.45f -> Color(0xFFFFC107)
+                        else -> Color(0xFFFF9800)
+                    },
                     shadowElevation = 4.dp
                 ) {
                     Text(
-                        text = "${(photo.similarityScore * 100).toInt()}%",
+                        text = "${(photo.finalScore * 100).toInt()}%",
                         modifier = Modifier.padding(horizontal = 8.dp, vertical = 4.dp),
                         style = MaterialTheme.typography.labelSmall,
                         fontWeight = FontWeight.Bold,
-                        color = MaterialTheme.colorScheme.onPrimary
+                        color = Color.White
                     )
                 }
             }
@@ -332,7 +389,7 @@ private fun PhotoCard(
             }
 
             // Face count badge (bottom-right)
-            if (photo.faceCount > 1) {
+            if (photo.faceCount > 1 && !isNegative) {
                 Surface(
                     modifier = Modifier
                         .align(Alignment.BottomEnd)
@@ -395,6 +452,7 @@ private fun RollingScanBottomBar(
     isReadyForTraining: Boolean,
     validationMessage: String?,
     onSelectTopN: (Int) -> Unit,
+    onSelectAboveThreshold: (Float) -> Unit,
     onSubmit: () -> Unit
 ) {
     Surface(
@@ -416,30 +474,41 @@ private fun RollingScanBottomBar(
                 )
             }
 
+            // First row: threshold selection
             Row(
                 modifier = Modifier.fillMaxWidth(),
-                horizontalArrangement = Arrangement.spacedBy(8.dp)
+                horizontalArrangement = Arrangement.spacedBy(6.dp)
             ) {
-                // Quick select buttons
                 OutlinedButton(
-                    onClick = { onSelectTopN(10) },
-                    modifier = Modifier.weight(1f)
+                    onClick = { onSelectAboveThreshold(0.60f) },
+                    modifier = Modifier.weight(1f),
+                    contentPadding = PaddingValues(horizontal = 8.dp, vertical = 4.dp)
                 ) {
-                    Text("Top 10")
+                    Text(">60%", style = MaterialTheme.typography.labelSmall)
+                }
+                OutlinedButton(
+                    onClick = { onSelectAboveThreshold(0.50f) },
+                    modifier = Modifier.weight(1f),
+                    contentPadding = PaddingValues(horizontal = 8.dp, vertical = 4.dp)
+                ) {
+                    Text(">50%", style = MaterialTheme.typography.labelSmall)
+                }
+                OutlinedButton(
+                    onClick = { onSelectTopN(15) },
+                    modifier = Modifier.weight(1f),
+                    contentPadding = PaddingValues(horizontal = 8.dp, vertical = 4.dp)
+                ) {
+                    Text("Top 15", style = MaterialTheme.typography.labelSmall)
+                }
             }
 
-                OutlinedButton(
-                    onClick = { onSelectTopN(20) },
-                    modifier = Modifier.weight(1f)
-                ) {
-                    Text("Top 20")
-                }
+            Spacer(Modifier.height(8.dp))
 
-                // Submit button
+            // Second row: submit
             Button(
                 onClick = onSubmit,
                 enabled = isReadyForTraining,
-                    modifier = Modifier.weight(1.5f)
+                modifier = Modifier.fillMaxWidth()
             ) {
                 Icon(
                     Icons.Default.Done,
@@ -447,8 +516,7 @@ private fun RollingScanBottomBar(
                     modifier = Modifier.size(18.dp)
                 )
                 Spacer(Modifier.width(8.dp))
-                    Text("Train ($selectedCount)")
-                }
+                Text("Train Model ($selectedCount photos)")
             }
         }
     }

app/src/main/java/com/placeholder/sherpai2/ui/rollingscan/Rollingscanviewmodel.kt

@@ -44,6 +44,11 @@ class RollingScanViewModel @Inject constructor(
         private const val TAG = "RollingScanVM"
         private const val DEBOUNCE_DELAY_MS = 300L
         private const val MIN_PHOTOS_FOR_TRAINING = 15
+
+        // Progressive thresholds based on selection count
+        private const val FLOOR_FEW_SEEDS = 0.30f      // 1-3 seeds
+        private const val FLOOR_MEDIUM_SEEDS = 0.40f   // 4-10 seeds
+        private const val FLOOR_MANY_SEEDS = 0.50f     // 10+ seeds
     }
 
     // ═══════════════════════════════════════════════════════════
@@ -71,6 +76,11 @@ class RollingScanViewModel @Inject constructor(
     // Cache of selected embeddings
     private val selectedEmbeddings = mutableListOf<FloatArray>()
 
+    // Negative embeddings (marked as "not this person")
+    private val _negativeImageIds = MutableStateFlow<Set<String>>(emptySet())
+    val negativeImageIds: StateFlow<Set<String>> = _negativeImageIds.asStateFlow()
+    private val negativeEmbeddings = mutableListOf<FloatArray>()
+
     // All available image IDs
     private var allImageIds: List<String> = emptyList()
 
@@ -156,24 +166,55 @@ class RollingScanViewModel @Inject constructor(
             current.remove(imageId)
 
             viewModelScope.launch {
-                // Remove embedding from cache
                 val cached = faceCacheDao.getEmbeddingByImageId(imageId)
                 cached?.getEmbedding()?.let { selectedEmbeddings.remove(it) }
             }
         } else {
-            // Select
+            // Select (and remove from negatives if present)
             current.add(imageId)
+            if (imageId in _negativeImageIds.value) {
+                toggleNegative(imageId)
+            }
 
             viewModelScope.launch {
-                // Add embedding to cache
                 val cached = faceCacheDao.getEmbeddingByImageId(imageId)
                 cached?.getEmbedding()?.let { selectedEmbeddings.add(it) }
             }
         }
 
-        _selectedImageIds.value = current
+        _selectedImageIds.value = current.toSet()  // Immutable copy
+
+        scanDebouncer.debounce {
+            triggerRollingScan()
+        }
+    }
+
+    /**
+     * Toggle negative marking ("Not this person")
+     */
+    fun toggleNegative(imageId: String) {
+        val current = _negativeImageIds.value.toMutableSet()
+
+        if (imageId in current) {
+            current.remove(imageId)
+            viewModelScope.launch {
+                val cached = faceCacheDao.getEmbeddingByImageId(imageId)
+                cached?.getEmbedding()?.let { negativeEmbeddings.remove(it) }
+            }
+        } else {
+            current.add(imageId)
+            // Remove from selected if present
+            if (imageId in _selectedImageIds.value) {
+                toggleSelection(imageId)
+            }
+            viewModelScope.launch {
+                val cached = faceCacheDao.getEmbeddingByImageId(imageId)
+                cached?.getEmbedding()?.let { negativeEmbeddings.add(it) }
+            }
+        }
+
+        _negativeImageIds.value = current.toSet()  // Immutable copy
 
-        // Debounced rescan
         scanDebouncer.debounce {
             triggerRollingScan()
         }
@@ -190,13 +231,33 @@ class RollingScanViewModel @Inject constructor(
 
         val current = _selectedImageIds.value.toMutableSet()
         current.addAll(topPhotos)
-        _selectedImageIds.value = current
+        _selectedImageIds.value = current.toSet()  // Immutable copy
 
         viewModelScope.launch {
-            // Add embeddings
             val embeddings = faceCacheDao.getEmbeddingsForImages(topPhotos.toList())
             selectedEmbeddings.addAll(embeddings.mapNotNull { it.getEmbedding() })
+            triggerRollingScan()
+        }
+    }
 
+    /**
+     * Select all photos above a similarity threshold
+     */
+    fun selectAllAboveThreshold(threshold: Float) {
+        val photosAbove = _rankedPhotos.value
+            .filter { it.finalScore >= threshold }
+            .map { it.imageId }
+
+        val current = _selectedImageIds.value.toMutableSet()
+        current.addAll(photosAbove)
+        _selectedImageIds.value = current.toSet()  // Immutable copy
+
+        viewModelScope.launch {
+            val newIds = photosAbove.filter { it !in _selectedImageIds.value }
+            if (newIds.isNotEmpty()) {
+                val embeddings = faceCacheDao.getEmbeddingsForImages(newIds)
+                selectedEmbeddings.addAll(embeddings.mapNotNull { it.getEmbedding() })
+            }
             triggerRollingScan()
         }
     }
@@ -207,17 +268,24 @@ class RollingScanViewModel @Inject constructor(
     fun clearSelection() {
         _selectedImageIds.value = emptySet()
         selectedEmbeddings.clear()
-
-        // Reset ranking
         _rankedPhotos.value = emptyList()
     }
 
+    /**
+     * Clear negative markings
+     */
+    fun clearNegatives() {
+        _negativeImageIds.value = emptySet()
+        negativeEmbeddings.clear()
+        scanDebouncer.debounce { triggerRollingScan() }
+    }
+
     // ═══════════════════════════════════════════════════════════
     // ROLLING SCAN LOGIC
     // ═══════════════════════════════════════════════════════════
 
     /**
-     * CORE: Trigger rolling similarity scan
+     * CORE: Trigger rolling similarity scan with progressive filtering
      */
     private suspend fun triggerRollingScan() {
         if (selectedEmbeddings.isEmpty()) {
@@ -228,7 +296,15 @@ class RollingScanViewModel @Inject constructor(
         try {
             _isScanning.value = true
 
-            Log.d(TAG, "Starting scan with ${selectedEmbeddings.size} selected embeddings")
+            val selectionCount = selectedEmbeddings.size
+            Log.d(TAG, "Starting scan with $selectionCount selected, ${negativeEmbeddings.size} negative")
+
+            // Progressive threshold based on selection count
+            val similarityFloor = when {
+                selectionCount <= 3 -> FLOOR_FEW_SEEDS
+                selectionCount <= 10 -> FLOOR_MEDIUM_SEEDS
+                else -> FLOOR_MANY_SEEDS
+            }
 
             // Calculate centroid from selected embeddings
             val centroid = faceSimilarityScorer.calculateCentroid(selectedEmbeddings)
@@ -240,17 +316,38 @@ class RollingScanViewModel @Inject constructor(
                 centroid = centroid
             )
 
-            // Update image URIs in scored photos
-            val photosWithUris = scoredPhotos.map { photo ->
+            // Apply negative penalty, quality boost, and floor filter
+            val filteredPhotos = scoredPhotos
+                .map { photo ->
+                    // Calculate max similarity to any negative embedding
+                    val negativePenalty = if (negativeEmbeddings.isNotEmpty()) {
+                        negativeEmbeddings.maxOfOrNull { neg ->
+                            cosineSimilarity(photo.cachedEmbedding, neg)
+                        } ?: 0f
+                    } else 0f
+
+                    // Quality multiplier: solo face, large face, good quality
+                    val qualityMultiplier = 1f +
+                        (if (photo.faceCount == 1) 0.15f else 0f) +
+                        (if (photo.faceAreaRatio > 0.15f) 0.10f else 0f) +
+                        (if (photo.qualityScore > 0.7f) 0.10f else 0f)
+
+                    // Final score = (similarity - negativePenalty) * qualityMultiplier
+                    val adjustedScore = ((photo.similarityScore - negativePenalty * 0.5f) * qualityMultiplier)
+                        .coerceIn(0f, 1f)
+
                     photo.copy(
-                    imageUri = imageUriCache[photo.imageId] ?: photo.imageId
+                        imageUri = imageUriCache[photo.imageId] ?: photo.imageId,
+                        finalScore = adjustedScore
                     )
                 }
+                .filter { it.finalScore >= similarityFloor }  // Apply floor
+                .filter { it.imageId !in _negativeImageIds.value }  // Hide negatives
+                .sortedByDescending { it.finalScore }
 
-            Log.d(TAG, "Scan complete. Scored ${photosWithUris.size} photos")
+            Log.d(TAG, "Scan complete. ${filteredPhotos.size} photos above floor $similarityFloor")
 
-            // Update ranked list
-            _rankedPhotos.value = photosWithUris
+            _rankedPhotos.value = filteredPhotos
 
         } catch (e: Exception) {
             Log.e(TAG, "Scan failed", e)
@@ -259,6 +356,19 @@ class RollingScanViewModel @Inject constructor(
         }
     }
 
+    private fun cosineSimilarity(a: FloatArray, b: FloatArray): Float {
+        if (a.size != b.size) return 0f
+        var dot = 0f
+        var normA = 0f
+        var normB = 0f
+        for (i in a.indices) {
+            dot += a[i] * b[i]
+            normA += a[i] * a[i]
+            normB += b[i] * b[i]
+        }
+        return if (normA > 0 && normB > 0) dot / (kotlin.math.sqrt(normA) * kotlin.math.sqrt(normB)) else 0f
+    }
+
     // ═══════════════════════════════════════════════════════════
     // SUBMISSION
     // ═══════════════════════════════════════════════════════════
@@ -299,9 +409,11 @@ class RollingScanViewModel @Inject constructor(
     fun reset() {
         _uiState.value = RollingScanState.Idle
         _selectedImageIds.value = emptySet()
+        _negativeImageIds.value = emptySet()
         _rankedPhotos.value = emptyList()
         _isScanning.value = false
         selectedEmbeddings.clear()
+        negativeEmbeddings.clear()
         allImageIds = emptyList()
         imageUriCache = emptyMap()
         scanDebouncer.cancel()

app/src/main/java/com/placeholder/sherpai2/ui/trainingprep/Facedetectionhelper.kt

@@ -6,8 +6,11 @@ import android.graphics.BitmapFactory
 import android.graphics.Rect
 import android.net.Uri
 import com.google.mlkit.vision.common.InputImage
+import com.google.mlkit.vision.face.Face
 import com.google.mlkit.vision.face.FaceDetection
 import com.google.mlkit.vision.face.FaceDetectorOptions
+import com.placeholder.sherpai2.domain.clustering.FaceQualityFilter
+import com.placeholder.sherpai2.ml.FaceNormalizer
 import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.async
 import kotlinx.coroutines.awaitAll
@@ -64,21 +67,30 @@ class FaceDetectionHelper(private val context: Context) {
                 val inputImage = InputImage.fromBitmap(bitmap, 0)
                 val faces = detector.process(inputImage).await()
 
-                // Sort by face size (area) to get the largest face
-                val sortedFaces = faces.sortedByDescending { face ->
+                // Filter to quality faces - use lenient scanning filter
+                // (Discovery filter was too strict, rejecting faces from rolling scan)
+                val qualityFaces = faces.filter { face ->
+                    FaceQualityFilter.validateForScanning(
+                        face = face,
+                        imageWidth = bitmap.width,
+                        imageHeight = bitmap.height
+                    )
+                }
+
+                // Sort by face size (area) to get the largest quality face
+                val sortedFaces = qualityFaces.sortedByDescending { face ->
                     face.boundingBox.width() * face.boundingBox.height()
                 }
 
                 val croppedFace = if (sortedFaces.isNotEmpty()) {
-                    // Crop the LARGEST detected face (most likely the subject)
-                    cropFaceFromBitmap(bitmap, sortedFaces[0].boundingBox)
+                    FaceNormalizer.cropAndNormalize(bitmap, sortedFaces[0])
                 } else null
 
                 FaceDetectionResult(
                     uri = uri,
-                    hasFace = faces.isNotEmpty(),
-                    faceCount = faces.size,
-                    faceBounds = faces.map { it.boundingBox },
+                    hasFace = qualityFaces.isNotEmpty(),
+                    faceCount = qualityFaces.size,
+                    faceBounds = qualityFaces.map { it.boundingBox },
                     croppedFaceBitmap = croppedFace
                 )
             } catch (e: Exception) {

app/src/main/java/com/placeholder/sherpai2/ui/trainingprep/ScanResultsScreen.kt

@@ -51,21 +51,8 @@ fun ScanResultsScreen(
         }
     }
 
-    Scaffold(
-        topBar = {
-            TopAppBar(
-                title = { Text("Train New Person") },
-                colors = TopAppBarDefaults.topAppBarColors(
-                    containerColor = MaterialTheme.colorScheme.primaryContainer
-                )
-            )
-        }
-    ) { paddingValues ->
-        Box(
-            modifier = Modifier
-                .fillMaxSize()
-                .padding(paddingValues)
-        ) {
+    // No Scaffold - MainScreen provides TopAppBar
+    Box(modifier = Modifier.fillMaxSize()) {
         when (state) {
             is ScanningState.Idle -> {}
 
@@ -77,8 +64,6 @@ fun ScanResultsScreen(
                 ImprovedResultsView(
                     result = state.sanityCheckResult,
                     onContinue = {
-                            // PersonInfo already captured in TrainingScreen!
-                            // Just start training with stored info
                         trainViewModel.createFaceModel(
                             trainViewModel.getPersonInfo()?.name ?: "Unknown"
                         )
@@ -103,7 +88,6 @@ fun ScanResultsScreen(
             TrainingOverlay(trainingState = trainingState as TrainingState.Processing)
         }
     }
-    }
 
     showFacePickerDialog?.let { result ->
         FacePickerDialog(

app/src/main/java/com/placeholder/sherpai2/ui/trainingprep/TrainViewModel.kt

@@ -5,11 +5,18 @@ import android.graphics.Bitmap
 import android.net.Uri
 import androidx.lifecycle.AndroidViewModel
 import androidx.lifecycle.viewModelScope
+import androidx.datastore.preferences.core.booleanPreferencesKey
+import androidx.datastore.preferences.preferencesDataStore
+import androidx.work.WorkManager
+import android.content.Context
 import com.placeholder.sherpai2.data.local.entity.PersonEntity
 import com.placeholder.sherpai2.data.repository.FaceRecognitionRepository
 import com.placeholder.sherpai2.ml.FaceNetModel
+import com.placeholder.sherpai2.workers.LibraryScanWorker
 import dagger.hilt.android.lifecycle.HiltViewModel
 import kotlinx.coroutines.flow.MutableStateFlow
+import kotlinx.coroutines.flow.first
+import kotlinx.coroutines.flow.map
 import kotlinx.coroutines.flow.StateFlow
 import kotlinx.coroutines.flow.asStateFlow
 import kotlinx.coroutines.launch
@@ -48,15 +55,20 @@ data class PersonInfo(
 /**
  * FIXED TrainViewModel with proper exclude functionality and efficient replace
  */
+private val android.content.Context.dataStore by preferencesDataStore(name = "settings")
+private val KEY_BACKGROUND_TAGGING = booleanPreferencesKey("background_recognition_tagging")
+
 @HiltViewModel
 class TrainViewModel @Inject constructor(
     application: Application,
     private val faceRecognitionRepository: FaceRecognitionRepository,
-    private val faceNetModel: FaceNetModel
+    private val faceNetModel: FaceNetModel,
+    private val workManager: WorkManager
 ) : AndroidViewModel(application) {
 
     private val sanityChecker = TrainingSanityChecker(application)
     private val faceDetectionHelper = FaceDetectionHelper(application)
+    private val dataStore = application.dataStore
 
     private val _uiState = MutableStateFlow<ScanningState>(ScanningState.Idle)
     val uiState: StateFlow<ScanningState> = _uiState.asStateFlow()
@@ -174,6 +186,20 @@ class TrainViewModel @Inject constructor(
                     relationship = person.relationship
                 )
 
+                // Trigger library scan if setting enabled
+                val backgroundTaggingEnabled = dataStore.data
+                    .map { it[KEY_BACKGROUND_TAGGING] ?: true }
+                    .first()
+
+                if (backgroundTaggingEnabled) {
+                    // Use default threshold (0.62 solo, 0.68 group)
+                    val scanRequest = LibraryScanWorker.createWorkRequest(
+                        personId = personId,
+                        personName = personName
+                    )
+                    workManager.enqueue(scanRequest)
+                }
+
             } catch (e: Exception) {
                 _trainingState.value = TrainingState.Error(
                     e.message ?: "Failed to create face model"
@@ -355,7 +381,7 @@ class TrainViewModel @Inject constructor(
             faceDetectionResults = updatedFaceResults,
             validationErrors = updatedErrors,
             validImagesWithFaces = updatedValidImages,
-            excludedImages = excludedImages
+            excludedImages = excludedImages.toSet()  // Immutable copy for Compose state detection
         )
     }
 

app/src/main/java/com/placeholder/sherpai2/ui/trainingprep/Trainingphotoselectorscreen.kt

@@ -49,6 +49,7 @@ fun TrainingPhotoSelectorScreen(
     val isRanking by viewModel.isRanking.collectAsStateWithLifecycle()
     val showPremiumOnly by viewModel.showPremiumOnly.collectAsStateWithLifecycle()
     val premiumCount by viewModel.premiumCount.collectAsStateWithLifecycle()
+    val embeddingProgress by viewModel.embeddingProgress.collectAsStateWithLifecycle()
 
     Scaffold(
         topBar = {
@@ -154,8 +155,34 @@ fun TrainingPhotoSelectorScreen(
                     Box(
                         modifier = Modifier.fillMaxSize(),
                         contentAlignment = Alignment.Center
+                    ) {
+                        Column(
+                            horizontalAlignment = Alignment.CenterHorizontally,
+                            verticalArrangement = Arrangement.spacedBy(16.dp)
                         ) {
                             CircularProgressIndicator()
+                            // Capture value to avoid race condition
+                            val progress = embeddingProgress
+                            if (progress != null) {
+                                Text(
+                                    "Preparing faces: ${progress.current}/${progress.total}",
+                                    style = MaterialTheme.typography.bodyMedium,
+                                    color = MaterialTheme.colorScheme.onSurfaceVariant
+                                )
+                                LinearProgressIndicator(
+                                    progress = { progress.current.toFloat() / progress.total },
+                                    modifier = Modifier
+                                        .width(200.dp)
+                                        .padding(top = 8.dp)
+                                )
+                            } else {
+                                Text(
+                                    "Loading premium faces...",
+                                    style = MaterialTheme.typography.bodyMedium,
+                                    color = MaterialTheme.colorScheme.onSurfaceVariant
+                                )
+                            }
+                        }
                     }
                 }
                 photos.isEmpty() -> {

app/src/main/java/com/placeholder/sherpai2/ui/trainingprep/Trainingphotoselectorviewmodel.kt

@@ -1,20 +1,31 @@
 package com.placeholder.sherpai2.ui.trainingprep
 
+import android.app.Application
+import android.graphics.Bitmap
+import android.graphics.BitmapFactory
+import android.graphics.Rect
+import android.net.Uri
 import android.util.Log
-import androidx.lifecycle.ViewModel
+import androidx.lifecycle.AndroidViewModel
 import androidx.lifecycle.viewModelScope
 import com.placeholder.sherpai2.data.local.dao.FaceCacheDao
 import com.placeholder.sherpai2.data.local.dao.ImageDao
+import com.placeholder.sherpai2.data.local.entity.FaceCacheEntity
 import com.placeholder.sherpai2.data.local.entity.ImageEntity
 import com.placeholder.sherpai2.domain.similarity.FaceSimilarityScorer
+import com.placeholder.sherpai2.ml.FaceNetModel
 import dagger.hilt.android.lifecycle.HiltViewModel
+import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.Job
 import kotlinx.coroutines.delay
 import kotlinx.coroutines.flow.MutableStateFlow
 import kotlinx.coroutines.flow.StateFlow
 import kotlinx.coroutines.flow.asStateFlow
 import kotlinx.coroutines.launch
+import kotlinx.coroutines.withContext
 import javax.inject.Inject
+import kotlin.math.max
+import kotlin.math.min
 
 /**
  * TrainingPhotoSelectorViewModel - PREMIUM GRID + ROLLING SCAN
@@ -27,15 +38,18 @@ import javax.inject.Inject
  */
 @HiltViewModel
 class TrainingPhotoSelectorViewModel @Inject constructor(
+    application: Application,
     private val imageDao: ImageDao,
     private val faceCacheDao: FaceCacheDao,
-    private val faceSimilarityScorer: FaceSimilarityScorer
-) : ViewModel() {
+    private val faceSimilarityScorer: FaceSimilarityScorer,
+    private val faceNetModel: FaceNetModel
+) : AndroidViewModel(application) {
 
     companion object {
         private const val TAG = "PremiumSelector"
         private const val MIN_SEEDS_FOR_ROLLING_SCAN = 1
         private const val MAX_SEEDS_FOR_ROLLING_SCAN = 5
+        private const val MAX_EMBEDDINGS_TO_GENERATE = 500
     }
 
     // All photos (for fallback / full list)
@@ -56,6 +70,12 @@ class TrainingPhotoSelectorViewModel @Inject constructor(
     private val _isRanking = MutableStateFlow(false)
     val isRanking: StateFlow<Boolean> = _isRanking.asStateFlow()
 
+    // Embedding generation progress
+    private val _embeddingProgress = MutableStateFlow<EmbeddingProgress?>(null)
+    val embeddingProgress: StateFlow<EmbeddingProgress?> = _embeddingProgress.asStateFlow()
+
+    data class EmbeddingProgress(val current: Int, val total: Int)
+
     // Premium mode toggle
     private val _showPremiumOnly = MutableStateFlow(true)
     val showPremiumOnly: StateFlow<Boolean> = _showPremiumOnly.asStateFlow()
@@ -79,20 +99,47 @@ class TrainingPhotoSelectorViewModel @Inject constructor(
 
     /**
      * Load PREMIUM faces first (solo, large, frontal, high quality)
+     * If no embeddings exist, generate them on-demand for premium candidates
      */
     private fun loadPremiumFaces() {
         viewModelScope.launch {
             try {
                 _isLoading.value = true
 
-                // Get premium faces from cache
-                val premiumFaceCache = faceCacheDao.getPremiumFaces(
+                // First check if premium faces with embeddings exist
+                var premiumFaceCache = faceCacheDao.getPremiumFaces(
                     minAreaRatio = 0.10f,
                     minQuality = 0.7f,
                     limit = 500
                 )
 
-                Log.d(TAG, "✅ Found ${premiumFaceCache.size} premium faces")
+                Log.d(TAG, "📊 Found ${premiumFaceCache.size} premium faces with embeddings")
+
+                // If no premium faces with embeddings, generate them on-demand
+                if (premiumFaceCache.isEmpty()) {
+                    Log.d(TAG, "⚠️ No premium faces with embeddings - generating on-demand")
+
+                    val candidates = faceCacheDao.getPremiumFaceCandidatesNeedingEmbeddings(
+                        minAreaRatio = 0.10f,
+                        minQuality = 0.7f,
+                        limit = MAX_EMBEDDINGS_TO_GENERATE
+                    )
+
+                    Log.d(TAG, "📦 Found ${candidates.size} premium candidates needing embeddings")
+
+                    if (candidates.isNotEmpty()) {
+                        generateEmbeddingsForCandidates(candidates)
+
+                        // Re-query after generating
+                        premiumFaceCache = faceCacheDao.getPremiumFaces(
+                            minAreaRatio = 0.10f,
+                            minQuality = 0.7f,
+                            limit = 500
+                        )
+                        Log.d(TAG, "✅ After generation: ${premiumFaceCache.size} premium faces")
+                    }
+                }
+
                 _premiumCount.value = premiumFaceCache.size
 
                 // Get corresponding ImageEntities
@@ -117,10 +164,108 @@ class TrainingPhotoSelectorViewModel @Inject constructor(
                 loadAllFaces()
             } finally {
                 _isLoading.value = false
+                _embeddingProgress.value = null
             }
         }
     }
 
+    /**
+     * Generate embeddings for premium face candidates
+     */
+    private suspend fun generateEmbeddingsForCandidates(candidates: List<FaceCacheEntity>) {
+        val context = getApplication<Application>()
+        val total = candidates.size
+        var processed = 0
+
+        withContext(Dispatchers.IO) {
+            // Get image URIs for candidates
+            val imageIds = candidates.map { it.imageId }.distinct()
+            val images = imageDao.getImagesByIds(imageIds)
+            val imageUriMap = images.associate { it.imageId to it.imageUri }
+
+            for (candidate in candidates) {
+                try {
+                    val imageUri = imageUriMap[candidate.imageId] ?: continue
+
+                    // Load bitmap
+                    val bitmap = loadBitmapOptimized(context, Uri.parse(imageUri)) ?: continue
+
+                    // Crop face
+                    val croppedFace = cropFaceWithPadding(bitmap, candidate.getBoundingBox())
+                    bitmap.recycle()
+
+                    if (croppedFace == null) continue
+
+                    // Generate embedding
+                    val embedding = faceNetModel.generateEmbedding(croppedFace)
+                    croppedFace.recycle()
+
+                    // Validate embedding
+                    if (embedding.any { it != 0f }) {
+                        // Save to database
+                        val embeddingJson = FaceCacheEntity.embeddingToJson(embedding)
+                        faceCacheDao.updateEmbedding(candidate.imageId, candidate.faceIndex, embeddingJson)
+                    }
+
+                } catch (e: Exception) {
+                    Log.w(TAG, "Failed to generate embedding for ${candidate.imageId}: ${e.message}")
+                }
+
+                processed++
+                withContext(Dispatchers.Main) {
+                    _embeddingProgress.value = EmbeddingProgress(processed, total)
+                }
+            }
+        }
+
+        Log.d(TAG, "✅ Generated embeddings for $processed/$total candidates")
+    }
+
+    private fun loadBitmapOptimized(context: android.content.Context, uri: Uri, maxDim: Int = 768): Bitmap? {
+        return try {
+            val options = BitmapFactory.Options().apply { inJustDecodeBounds = true }
+            context.contentResolver.openInputStream(uri)?.use { stream ->
+                BitmapFactory.decodeStream(stream, null, options)
+            }
+
+            var sampleSize = 1
+            while (options.outWidth / sampleSize > maxDim || options.outHeight / sampleSize > maxDim) {
+                sampleSize *= 2
+            }
+
+            val finalOptions = BitmapFactory.Options().apply {
+                inSampleSize = sampleSize
+                inPreferredConfig = Bitmap.Config.ARGB_8888
+            }
+
+            context.contentResolver.openInputStream(uri)?.use { stream ->
+                BitmapFactory.decodeStream(stream, null, finalOptions)
+            }
+        } catch (e: Exception) {
+            Log.w(TAG, "Failed to load bitmap: ${e.message}")
+            null
+        }
+    }
+
+    private fun cropFaceWithPadding(bitmap: Bitmap, boundingBox: Rect): Bitmap? {
+        return try {
+            val padding = (max(boundingBox.width(), boundingBox.height()) * 0.25f).toInt()
+            val left = max(0, boundingBox.left - padding)
+            val top = max(0, boundingBox.top - padding)
+            val right = min(bitmap.width, boundingBox.right + padding)
+            val bottom = min(bitmap.height, boundingBox.bottom + padding)
+            val width = right - left
+            val height = bottom - top
+
+            if (width > 0 && height > 0) {
+                Bitmap.createBitmap(bitmap, left, top, width, height)
+            } else null
+        } catch (e: Exception) {
+            Log.w(TAG, "Failed to crop face: ${e.message}")
+            null
+        }
+    }
+
     /**
      * Fallback: load all photos with faces
      */

app/src/main/java/com/placeholder/sherpai2/workers/Libraryscanworker.kt

@@ -9,6 +9,9 @@ import com.google.mlkit.vision.common.InputImage
 import com.google.mlkit.vision.face.FaceDetection
 import com.google.mlkit.vision.face.FaceDetectorOptions
 import com.placeholder.sherpai2.data.local.dao.FaceModelDao
+import com.placeholder.sherpai2.data.local.dao.PersonDao
+import com.placeholder.sherpai2.domain.clustering.FaceQualityFilter
+import com.placeholder.sherpai2.ml.FaceNormalizer
 import com.placeholder.sherpai2.data.local.dao.ImageDao
 import com.placeholder.sherpai2.data.local.dao.PhotoFaceTagDao
 import com.placeholder.sherpai2.data.local.entity.PhotoFaceTagEntity
@@ -52,7 +55,8 @@ class LibraryScanWorker @AssistedInject constructor(
     @Assisted workerParams: WorkerParameters,
     private val imageDao: ImageDao,
     private val faceModelDao: FaceModelDao,
-    private val photoFaceTagDao: PhotoFaceTagDao
+    private val photoFaceTagDao: PhotoFaceTagDao,
+    private val personDao: PersonDao
 ) : CoroutineWorker(context, workerParams) {
 
     companion object {
@@ -65,7 +69,8 @@ class LibraryScanWorker @AssistedInject constructor(
         const val KEY_MATCHES_FOUND = "matches_found"
         const val KEY_PHOTOS_SCANNED = "photos_scanned"
 
-        private const val DEFAULT_THRESHOLD = 0.70f  // Slightly looser than validation
+        private const val DEFAULT_THRESHOLD = 0.62f  // Solo photos
+        private const val GROUP_THRESHOLD = 0.68f   // Group photos (stricter)
         private const val BATCH_SIZE = 20
         private const val MAX_RETRIES = 3
 
@@ -137,16 +142,40 @@ class LibraryScanWorker @AssistedInject constructor(
                 )
             }
 
+            // Step 2.5: Load person to check isChild flag
+            val person = withContext(Dispatchers.IO) {
+                personDao.getPersonById(personId)
+            }
+            val isChildTarget = person?.isChild ?: false
+
             // Step 3: Initialize ML components
             val faceNetModel = FaceNetModel(context)
             val detector = FaceDetection.getClient(
                 FaceDetectorOptions.Builder()
                     .setPerformanceMode(FaceDetectorOptions.PERFORMANCE_MODE_ACCURATE)
+                    .setLandmarkMode(FaceDetectorOptions.LANDMARK_MODE_ALL)  // Needed for age estimation
                     .setMinFaceSize(0.15f)
                     .build()
             )
 
-            val modelEmbedding = faceModel.getEmbeddingArray()
+            // Distribution-based minimum threshold (self-calibrating)
+            val distributionMin = (faceModel.averageConfidence - 2 * faceModel.similarityStdDev)
+                .coerceAtLeast(faceModel.similarityMin - 0.05f)
+                .coerceAtLeast(0.50f)  // Never go below 0.50 absolute floor
+
+            // Get ALL centroids for multi-centroid matching (critical for children)
+            val modelCentroids = faceModel.getCentroids().map { it.getEmbeddingArray() }
+            if (modelCentroids.isEmpty()) {
+                return@withContext Result.failure(workDataOf("error" to "No centroids in model"))
+            }
+
+            // Load ALL other models for "best match wins" comparison
+            // This prevents tagging siblings incorrectly
+            val allModels = withContext(Dispatchers.IO) { faceModelDao.getAllActiveFaceModels() }
+            val otherModelCentroids = allModels
+                .filter { it.id != faceModel.id }
+                .map { model -> model.id to model.getCentroids().map { it.getEmbeddingArray() } }
+
             var matchesFound = 0
             var photosScanned = 0
 
@@ -164,10 +193,13 @@ class LibraryScanWorker @AssistedInject constructor(
                                 photo = photo,
                                 personId = personId,
                                 faceModelId = faceModel.id,
-                                modelEmbedding = modelEmbedding,
+                                modelCentroids = modelCentroids,
+                                otherModelCentroids = otherModelCentroids,
                                 faceNetModel = faceNetModel,
                                 detector = detector,
-                                threshold = threshold
+                                threshold = threshold,
+                                distributionMin = distributionMin,
+                                isChildTarget = isChildTarget
                             )
 
                             if (tags.isNotEmpty()) {
@@ -228,10 +260,13 @@ class LibraryScanWorker @AssistedInject constructor(
         photo: com.placeholder.sherpai2.data.local.entity.ImageEntity,
         personId: String,
         faceModelId: String,
-        modelEmbedding: FloatArray,
+        modelCentroids: List<FloatArray>,
+        otherModelCentroids: List<Pair<String, List<FloatArray>>>,
         faceNetModel: FaceNetModel,
         detector: com.google.mlkit.vision.face.FaceDetector,
-        threshold: Float
+        threshold: Float,
+        distributionMin: Float,
+        isChildTarget: Boolean
     ): List<PhotoFaceTagEntity> = withContext(Dispatchers.IO) {
 
         try {
@@ -243,43 +278,94 @@ class LibraryScanWorker @AssistedInject constructor(
             val inputImage = InputImage.fromBitmap(bitmap, 0)
             val faces = detector.process(inputImage).await()
 
-            // Check each face
-            val tags = faces.mapNotNull { face ->
+            if (faces.isEmpty()) {
+                bitmap.recycle()
+                return@withContext emptyList()
+            }
+
+            // Use higher threshold for group photos
+            val isGroupPhoto = faces.size > 1
+            val effectiveThreshold = if (isGroupPhoto) GROUP_THRESHOLD else threshold
+
+            // Track best match (only tag ONE face per image to avoid false positives)
+            var bestMatch: PhotoFaceTagEntity? = null
+            var bestSimilarity = 0f
+
+            // Check each face (filter by quality first)
+            for (face in faces) {
+                // Quality check
+                if (!FaceQualityFilter.validateForScanning(face, bitmap.width, bitmap.height)) {
+                    continue
+                }
+
+                // Skip very small faces
+                val faceArea = face.boundingBox.width() * face.boundingBox.height()
+                val imageArea = bitmap.width * bitmap.height
+                if (faceArea.toFloat() / imageArea < 0.02f) continue
+
+                // SIGNAL 2: Age plausibility check (if target is a child)
+                if (isChildTarget) {
+                    val ageGroup = FaceQualityFilter.estimateAgeGroup(face, bitmap.width, bitmap.height)
+                    if (ageGroup == FaceQualityFilter.AgeGroup.ADULT) {
+                        continue  // Reject clearly adult faces when searching for a child
+                    }
+                }
+
                 try {
-                    // Crop face
-                    val faceBitmap = android.graphics.Bitmap.createBitmap(
-                        bitmap,
-                        face.boundingBox.left.coerceIn(0, bitmap.width - 1),
-                        face.boundingBox.top.coerceIn(0, bitmap.height - 1),
-                        face.boundingBox.width().coerceAtMost(bitmap.width - face.boundingBox.left),
-                        face.boundingBox.height().coerceAtMost(bitmap.height - face.boundingBox.top)
-                    )
+                    // Crop and normalize face for best recognition
+                    val faceBitmap = FaceNormalizer.cropAndNormalize(bitmap, face)
+                        ?: continue
 
                     // Generate embedding
                     val faceEmbedding = faceNetModel.generateEmbedding(faceBitmap)
                     faceBitmap.recycle()
 
-                    // Calculate similarity
-                    val similarity = faceNetModel.calculateSimilarity(faceEmbedding, modelEmbedding)
+                    // Match against target person's centroids
+                    val targetSimilarity = modelCentroids.maxOfOrNull { centroid ->
+                        faceNetModel.calculateSimilarity(faceEmbedding, centroid)
+                    } ?: 0f
 
-                    if (similarity >= threshold) {
-                        PhotoFaceTagEntity.create(
+                    // SIGNAL 1: Distribution-based rejection
+                    // If similarity is below (mean - 2*stdDev) or (min - 0.05), it's a statistical outlier
+                    if (targetSimilarity < distributionMin) {
+                        continue  // Too far below training distribution
+                    }
+
+                    // SIGNAL 3: Basic threshold check
+                    if (targetSimilarity < effectiveThreshold) {
+                        continue
+                    }
+
+                    // SIGNAL 4: "Best match wins" - check if any OTHER model scores higher
+                    // This prevents tagging siblings incorrectly
+                    val bestOtherSimilarity = otherModelCentroids.maxOfOrNull { (_, centroids) ->
+                        centroids.maxOfOrNull { centroid ->
+                            faceNetModel.calculateSimilarity(faceEmbedding, centroid)
+                        } ?: 0f
+                    } ?: 0f
+
+                    val isTargetBestMatch = targetSimilarity > bestOtherSimilarity
+
+                    // All signals must pass
+                    if (isTargetBestMatch && targetSimilarity > bestSimilarity) {
+                        bestSimilarity = targetSimilarity
+                        bestMatch = PhotoFaceTagEntity.create(
                             imageId = photo.imageId,
                             faceModelId = faceModelId,
                             boundingBox = face.boundingBox,
-                            confidence = similarity,
+                            confidence = targetSimilarity,
                             faceEmbedding = faceEmbedding
                         )
-                    } else {
-                        null
                     }
                 } catch (e: Exception) {
-                    null
+                    // Skip this face
                 }
             }
 
             bitmap.recycle()
-            tags
+
+            // Return only the best match (or empty)
+            if (bestMatch != null) listOf(bestMatch) else emptyList()
 
         } catch (e: Exception) {
             emptyList()