Session
Interface for managing face recognition sessions. Provides functionality for face tracking, feature extraction, and analysis.
Methods
setTrackPreviewSize
Set the track preview size in the session.
setTrackPreviewSize(size: number): void
Parameters
| Name | Type | Description |
|---|---|---|
size | number | Size in pixels for tracking preview. Defaults to 192 |
Returns
void
setFaceDetectThreshold
Set the face detect threshold in the session.
setFaceDetectThreshold(threshold: number): void
Parameters
| Name | Type | Description |
|---|---|---|
threshold | number | Detection threshold value between 0 and 1 |
Returns
void
setFilterMinimumFacePixelSize
Set the minimum number of face pixels that the face detector can capture, and people below this number will be filtered.
setFilterMinimumFacePixelSize(size: number): void
Parameters
| Name | Type | Description |
|---|---|---|
size | number | Minimum size in pixels. Defaults to 0 |
Returns
void
setTrackModeSmoothRatio
Set the track mode smooth ratio in the session.
setTrackModeSmoothRatio(ratio: number): void
Parameters
| Name | Type | Description |
|---|---|---|
ratio | number | Smoothing ratio value between 0 and 1. Defaults to 0.05 |
Returns
void
setTrackModeNumSmoothCacheFrame
Set the track mode num smooth cache frame in the session.
setTrackModeNumSmoothCacheFrame(num: number): void
Parameters
| Name | Type | Description |
|---|---|---|
num | number | Number of frames to cache for smoothing. Defaults to 5 |
Returns
void
setTrackModeDetectInterval
Set the track model detect interval in the session.
setTrackModeDetectInterval(num: number): void
Parameters
| Name | Type | Description |
|---|---|---|
num | number | Interval value between detections. Defaults to 20 |
Returns
void
executeFaceTrack
Run face tracking in the session.
executeFaceTrack(imageStream: ImageStream): FaceData[]
Parameters
| Name | Type | Description |
|---|---|---|
imageStream | ImageStream | Input image stream to process |
Returns
FaceData[]– Array of detected face data objects.
extractFaceFeature
Extract a face feature from a given face.
extractFaceFeature(
imageStream: ImageStream,
faceToken: ArrayBuffer
): ArrayBuffer
Parameters
| Name | Type | Description |
|---|---|---|
imageStream | ImageStream | Input image stream to process |
faceToken | ArrayBuffer | Face token from previous detection |
Returns
ArrayBuffer– Face feature vector representing the detected face.
getFaceAlignmentImage
Get the face alignment image.
getFaceAlignmentImage(
imageStream: ImageStream,
faceToken: ArrayBuffer
): ImageBitmap
Parameters
| Name | Type | Description |
|---|---|---|
imageStream | ImageStream | Input image stream to process |
faceToken | ArrayBuffer | Face token from previous detection |
Returns
ImageBitmap– Aligned face image from the detection.
multipleFacePipelineProcess
Process multiple faces in a pipeline.
multipleFacePipelineProcess(
imageStream: ImageStream,
multipleFaceData: FaceData[],
parameter: SessionCustomParameter
): boolean
Parameters
| Name | Type | Description |
|---|---|---|
imageStream | ImageStream | Input image stream to process |
multipleFaceData | FaceData[] | Array of face data objects to process |
parameter | SessionCustomParameter | Configuration for feature enabling/disabling |
Returns
boolean– Returnstrueif the pipeline processing completed successfully; otherwisefalse.
getRGBLivenessConfidence
Get the RGB liveness confidence.
getRGBLivenessConfidence(): number[]
Returns
number[]– Confidence scores (0-1) per face.
getFaceQualityConfidence
Get the face quality predict confidence.
getFaceQualityConfidence(): number[]
Returns
number[]– Quality scores (0-1) per face.
getFaceMaskConfidence
Get the face mask confidence.
getFaceMaskConfidence(): number[]
Returns
number[]– Mask detection scores (0-1) per face.
getFaceInteractionState
Get the prediction results of face interaction.
getFaceInteractionState(): FaceInteractionState[]
Returns
FaceInteractionState[]– Array of face interaction states.
getFaceInteractionActionsResult
Get the prediction results of face interaction actions.
getFaceInteractionActionsResult(): FaceInteractionsAction[]
Returns
FaceInteractionsAction[]– Array of detected face interaction actions.
getFaceAttributeResult
Get the face attribute results.
getFaceAttributeResult(): FaceAttributeResult[]
Returns
FaceAttributeResult[]– Array of detected face attributes.
getFaceEmotionResult
Get the face emotion recognition results.
getFaceEmotionResult(): FaceEmotionResult[]
Returns
FaceEmotionResult[]- Array of detected face emotions
clearTrackingFace
Clear all currently tracked faces. Useful for resetting tracking state.
clearTrackingFace(): void
Returns
void
setTrackLostRecoveryMode
Set the track lost recovery mode (only for LightTrack mode).
setTrackLostRecoveryMode(enable: boolean): void
Parameters
| Name | Type | Description |
|---|---|---|
enable | boolean | Whether to enable track lost recovery (default: false) |
Returns
void
setLightTrackConfidenceThreshold
Set the light track confidence threshold (only for LightTrack mode).
setLightTrackConfidenceThreshold(value: number): void
Parameters
| Name | Type | Description |
|---|---|---|
value | number | Confidence threshold value (default: 0.1) |
Returns
void
reconfigure
Reconfigure the session with new parameters. Internally destroys and recreates the underlying session handle. The JS object reference remains stable.
reconfigure(
parameter: SessionCustomParameter,
detectMode: DetectMode,
maxDetectFaceNum: number,
detectPixelLevel: number,
trackByDetectModeFPS: number
): void
Parameters
| Name | Type | Description |
|---|---|---|
parameter | SessionCustomParameter | Custom parameters for the session |
detectMode | DetectMode | Face detection mode |
maxDetectFaceNum | number | Maximum number of faces to detect |
detectPixelLevel | number | Detection resolution level (-1 for default 320) |
trackByDetectModeFPS | number | Frame rate for tracking mode (-1 for default 30) |
Returns
void
getTrackPreviewSize
Get the current track preview size.
getTrackPreviewSize(): number
Returns
number- Current preview size in pixels
faceQualityDetect
Detect the quality of a single face without running the full pipeline. Useful for quick quality checks without enabling all pipeline features.
faceQualityDetect(faceToken: ArrayBuffer): number
Parameters
| Name | Type | Description |
|---|---|---|
faceToken | ArrayBuffer | Face token data |
Returns
number- Quality confidence score (0-1)
extractFaceFeatureFromAlignmentImage
Extract face features from an already-aligned face image. Use after calling getFaceAlignmentImage() to avoid redundant re-alignment.
extractFaceFeatureFromAlignmentImage(imageStream: ImageStream): ArrayBuffer
Parameters
| Name | Type | Description |
|---|---|---|
imageStream | ImageStream | Image stream of the aligned face |
Returns
ArrayBuffer- Extracted face feature vector