class · iOS 27.0+ · SDK 27.1
final class FaceGroupAnalyzerAn object that detects faces in images and groups them by person. ``FaceGroupAnalyzer`` is the central object in the face analysis pipeline. It detects faces in images, stores their data persistently in a working directory you choose, and clusters the faces into groups called *entities*, where each entity represents a distinct person. The analyzer stores all face data and cluster assignments in the directory you pass to ``FaceGroupAnalyzer/init(workingDirectory:)``. This directory persists between app launches, so subsequent runs resume from where the previous session ended. Call ``FaceGroupAnalyzer/purge(workingDirectory:)`` to remove all data from the directory. Use ``FaceGroupAnalyzer/identifyFaces(in:)`` to detect and match faces in images without storing anything. This is useful for recognizing people in new images against an existing gallery without modifying the analyzer's data.
fetchFacesIn(assetIDs: Array<MediaIntelligenceImageAsset_ID>): SDKHandlefinal func fetchFaces(in assetIDs: [MediaIntelligenceImageAsset.ID]) throws -> some AsyncSequence<(assetID: MediaIntelligenceImageAsset.ID, faces: Array<FaceGroupAnalyzer.Face>), any Error>
Returns the faces from the specified assets. - Parameter assetIDs: The asset identifiers to look up. - Returns: An async sequence of `(assetID, faces)` pairs.
fetchFacesFor(entityIDs: Array<FaceGroupAnalyzer_Entity_ID>): SDKHandlefinal func fetchFaces(for entityIDs: [FaceGroupAnalyzer.Entity.ID]) throws -> some AsyncSequence<(entityID: FaceGroupAnalyzer.Entity.ID, faces: Array<FaceGroupAnalyzer.Face>), any Error>
Returns the faces belonging to the specified entities. - Parameter entityIDs: The entity identifiers to look up. - Returns: An async sequence of `(entityID, faces)` pairs.
fetchFaces(faceIDs: Array<FaceGroupAnalyzer_Face_ID>): SDKHandlefinal func fetchFaces(_ faceIDs: [FaceGroupAnalyzer.Face.ID]) throws -> some AsyncSequence<FaceGroupAnalyzer.Face, any Error>
Returns the faces with the specified identifiers. - Parameter faceIDs: The face identifiers to look up. - Returns: An async sequence of matching ``FaceGroupAnalyzer/Face`` values.
get allAssetIDs(): SDKHandlefinal var allAssetIDs: some AsyncSequence<MediaIntelligenceImageAsset.ID, any Error> { get }An async sequence of all asset identifiers in the analyzer.
get allEntities(): SDKHandlefinal var allEntities: some AsyncSequence<FaceGroupAnalyzer.Entity, any Error> { get }An async sequence of all entities in the analyzer.
deleteAssets(assetIDs: Array<MediaIntelligenceImageAsset_ID>): Promise<void>final nonisolated(nonsending) func deleteAssets(_ assetIDs: [MediaIntelligenceImageAsset.ID]) async throwsRemoves the specified assets and their associated face data. - Parameter assetIDs: The identifiers of the assets to remove.
fetchAssetIDs(entityIDs: Array<FaceGroupAnalyzer_Entity_ID>): SDKHandlefinal func fetchAssetIDs(for entityIDs: [FaceGroupAnalyzer.Entity.ID]) throws -> some AsyncSequence<(entityID: FaceGroupAnalyzer.Entity.ID, assetIDs: Array<MediaIntelligenceImageAsset.ID>), any Error>
Returns the asset identifiers for the specified entities. - Parameter entityIDs: The entity identifiers to look up. - Returns: An async sequence of `(entityID, assetIDs)` pairs.
identifyFaces(assets: Array<MediaIntelligenceImageAsset>): Promise<SDKHandle>final nonisolated(nonsending) func identifyFaces(in assets: [MediaIntelligenceImageAsset]) async throws -> some AsyncSequence<(assetID: MediaIntelligenceImageAsset.ID, faces: Array<FaceGroupAnalyzer.Face>), any Error>
Detects and identifies faces without modifying the analyzer's data. Use this method to recognize people in new images against the existing gallery without storing any face data. Each returned face has an ``FaceGroupAnalyzer/Face/entityID`` if the framework matched it to a known entity, or `nil` if no match was found. - Parameter assets: The image assets to analyze. - Returns: An async sequence of `(assetID, faces)` pairs.
deleteAllAssets(): Promise<void>final nonisolated(nonsending) func deleteAllAssets() async throwsRemoves all assets and their associated face data from the analyzer.
get workingDirectory(): stringfinal let workingDirectory: URLA directory where the analyzer stores its face data and metadata.
static init(workingDirectory: string): FaceGroupAnalyzerinit(workingDirectory: URL) throwsCreates a face group analyzer at the specified directory. The analyzer loads any existing face data from `workingDirectory` on initialization, so subsequent runs resume from where the previous session ended. - Parameter workingDirectory: The directory where the analyzer stores its face data and metadata. The directory must already exist.
get allFacesByEntityID(): SDKHandlefinal var allFacesByEntityID: some AsyncSequence<(entityID: FaceGroupAnalyzer.Entity.ID, faces: Array<FaceGroupAnalyzer.Face>), any Error> { get }An async sequence of all faces, grouped by entity.
insertOrUpdateAssets(assets: Array<MediaIntelligenceImageAsset>): Promise<SDKHandle>final nonisolated(nonsending) func insertOrUpdateAssets(_ assets: [MediaIntelligenceImageAsset]) async throws -> some AsyncSequence<(assetID: MediaIntelligenceImageAsset.ID, faces: Array<FaceGroupAnalyzer.Face>), any Error>
Adds or replaces image assets in the analyzer. For each asset, the method detects faces and extracts facial data, then persists the results to the working directory. If an asset with the same identifier already exists, the method replaces its face data. After this call, ``FaceGroupAnalyzer/state`` becomes ``FaceGroupAnalyzer/State/stale`` until you call ``FaceGroupAnalyzer/update(subprogress:)``. Faces detected in this call have a `nil` ``FaceGroupAnalyzer/Face/entityID`` until clustering runs. - Parameter assets: The image assets to ingest. - Returns: An async sequence of `(assetID, faces)` pairs — one per asset — reporting the faces detected in each image.
get allAssetIDsByEntityID(): SDKHandlefinal var allAssetIDsByEntityID: some AsyncSequence<(entityID: FaceGroupAnalyzer.Entity.ID, assetIDs: Array<MediaIntelligenceImageAsset.ID>), any Error> { get }An async sequence of all asset identifiers, grouped by entity.
static purge(workingDirectory: string): Promise<void>nonisolated(nonsending) static func purge(workingDirectory: URL) async throwsRemoves all analyzer data from the specified directory. Call this method to permanently delete all face data, cluster assignments, and metadata from `workingDirectory`. This operation is irreversible. - Parameter workingDirectory: The directory to purge.
get state(): Promise<FaceGroupAnalyzer_State>nonisolated(nonsending) final var state: FaceGroupAnalyzer.State { get async }A value describing the current processing state of the analyzer.
get allFaces(): SDKHandlefinal var allFaces: some AsyncSequence<FaceGroupAnalyzer.Face, any Error> { get }An async sequence of all faces in the analyzer.
1 declarations were excluded from these bindings. Each reason comes from the generator or the iOS SDK compiler.
FaceGroupAnalyzer.update(subprogress:)
The iOS SDK compiler rejected this binding: class 'OneSDKCall' requires that 'Subprogress' conform to 'Copyable'