FaceTracker class

Tracks faces on consecutive images and automatically creates and updates associated face templates.

Implemented types

Constructors

FaceTracker()
Creates a new instance of the FaceTracker class.
FaceTracker.fromHandle(Pointer<id3FaceTracker> handle)
Creates a new instance of the FaceTracker class.

Properties

confidenceThreshold ↔ int
Confidence threshold, in the range 0;100. Hint: Default value is 50. Note: Setting a high threshold reduces false detection but can increase the number of undetected faces.
getter/setter pair
detectionModel ↔ FaceModel
Model used to detect and track faces. Hint: Default value is FaceDetector4B. Note: Some better accuracy/speed balances can be found by choosing another model.
getter/setter pair
encodingModel ↔ FaceModel
Model used to create features and assess consistancy among views of a given face. Hint: Default value is FaceEncoder9B. Some better accuracy/speed balances can be found by choosing another model.
getter/setter pair
handle → Pointer<id3FaceTracker>
Gets the native handle. return The native handle.
no setter
hashCode → int
The hash code for this object.
no setterinherited
matchThreshold ↔ int
Minimum match score to reach to preserve the ID of a tracked face between frame 't-1' and frame 't'. Hint: Default value is 3000 which corresponds to a False Match Rate of 1/1000.
getter/setter pair
maximumTrackedFaceAge ↔ int
Maximum number of consecutive non-detections to reach before deleting a tracked face. Hint: Default value is 30 which corresponds to 2s at a frame rate of 15 FPS. One must adapt this value to its needs in terms of tracker identity memory (in seconds) and measured frame rate on target platform.
getter/setter pair
minimumTrackedFaceAge ↔ int
Minimum number of consecutive detections to reach before creating a tracked face. Hint: Default value is 1 for FaceDetector4B since the false detection rate is low enough. If using a less accurate detector (such as FaceDetector3C) one might consider increasing a bit this value to avoid false tracks.
getter/setter pair
nmsIouThreshold ↔ int
Non-maximum suppression (NMS) intersection-over-union (IOU) threshold, in the range is 0;100. Hint: Default value is 40. Setting a high threshold allows to detect more overlapping faces which can be useful in a multi-face scenario. On the contrary, in a portrait scenario, a low NMS IOU threshold should be preferred.
getter/setter pair
runtimeType → Type
A representation of the runtime type of the object.
no setterinherited
threadCount ↔ int
Number of threads to be used for face detection and tracking. Hint: Default value is 1. Allocating more than one thread can increase the speed of the process.
getter/setter pair

Methods

detectFaces(Image image, TrackedFaceList trackedFaceList) → void
Detects faces in an image and update their info in a TrackedFaceList object. The algorithm searches for faces in the range (16px;512px). If the image is too large to fit this range, one must resize it before the tracking process. In a realtime process, one must use this function to keep the face IDs stable in time. Important: Loading a FaceDetector is required to use this function.
dispose() → void
Releases all resources used by this FaceTracker.
getConfidenceThreshold() → int
Gets the confidence threshold, in the range (0;100). Hint: Default value is 50. Note: Setting a high threshold reduces false detection but can increase the number of undetected faces.
getDetectionModel() → FaceModel
Gets the model used to detect and track faces. Hint: Default value is FaceDetector4B. Note: Some better accuracy/speed balances can be found by choosing another model.
getEncodingModel() → FaceModel
Gets the model used to create features and assess consistancy among views of a given face. Hint: Default value is FaceEncoder9B. Some better accuracy/speed balances can be found by choosing another model.
getMatchThreshold() → int
Gets the minimum match score to reach to preserve the ID of a tracked face between frame 't-1' and frame 't'. Hint: Default value is 3000 which corresponds to a False Match Rate of 1/1000.
getMaximumTrackedFaceAge() → int
Gets the maximum number of consecutive non-detections to reach before deleting a tracked face. Hint: Default value is 30 which corresponds to 2s at a frame rate of 15 FPS. One must adapt this value to its needs in terms of tracker identity memory (in seconds) and measured frame rate on target platform.
getMinimumTrackedFaceAge() → int
Gets the minimum number of consecutive detections to reach before creating a tracked face. Hint: Default value is 1 for FaceDetector4B since the false detection rate is low enough. If using a less accurate detector (such as FaceDetector3C) one might consider increasing a bit this value to avoid false tracks.
getNmsIouThreshold() → int
Gets the non-maximum suppression (NMS) intersection-over-union (IOU) threshold, in the range is (0;100). Hint: Default value is 40. Setting a high threshold allows to detect more overlapping faces which can be useful in a multi-face scenario. On the contrary, in a portrait scenario, a low NMS IOU threshold should be preferred.
getThreadCount() → int
Gets the number of threads to be used for face detection and tracking. Hint: Default value is 1. Allocating more than one thread can increase the speed of the process.
noSuchMethod(Invocation invocation) → dynamic
Invoked when a nonexistent method or property is accessed.
inherited
setConfidenceThreshold(int confidenceThreshold) → void
Sets the confidence threshold, in the range (0;100). Hint: Default value is 50. Note: Setting a high threshold reduces false detection but can increase the number of undetected faces.
setDetectionModel(FaceModel detectionModel) → void
Sets the model used to detect and track faces. Hint: Default value is FaceDetector4B. Note: Some better accuracy/speed balances can be found by choosing another model.
setEncodingModel(FaceModel encodingModel) → void
Sets the model used to create features and assess consistancy among views of a given face. Hint: Default value is FaceEncoder9B. Some better accuracy/speed balances can be found by choosing another model.
setMatchThreshold(int matchThreshold) → void
Sets the minimum match score to reach to preserve the ID of a tracked face between frame 't-1' and frame 't'. Hint: Default value is 3000 which corresponds to a False Match Rate of 1/1000.
setMaximumTrackedFaceAge(int maximumTrackedFaceAge) → void
Sets the maximum number of consecutive non-detections to reach before deleting a tracked face. Hint: Default value is 30 which corresponds to 2s at a frame rate of 15 FPS. One must adapt this value to its needs in terms of tracker identity memory (in seconds) and measured frame rate on target platform.
setMinimumTrackedFaceAge(int minimumTrackedFaceAge) → void
Sets the minimum number of consecutive detections to reach before creating a tracked face. Hint: Default value is 1 for FaceDetector4B since the false detection rate is low enough. If using a less accurate detector (such as FaceDetector3C) one might consider increasing a bit this value to avoid false tracks.
setNmsIouThreshold(int nmsIouThreshold) → void
Sets the non-maximum suppression (NMS) intersection-over-union (IOU) threshold, in the range is (0;100). Hint: Default value is 40. Setting a high threshold allows to detect more overlapping faces which can be useful in a multi-face scenario. On the contrary, in a portrait scenario, a low NMS IOU threshold should be preferred.
setThreadCount(int threadCount) → void
Sets the number of threads to be used for face detection and tracking. Hint: Default value is 1. Allocating more than one thread can increase the speed of the process.
toString() → String
A string representation of this object.
inherited
trackFaces(Image image, TrackedFaceList trackedFaceList) → void
Tracks faces in an image and update their info in a TrackedFaceList object. The algorithm searches for faces in the range (16px;512px). If the image is too large to fit this range, one must resize it before the tracking process. In a realtime process, one must use this function to keep the face IDs stable in time. Important: Loading a FaceDetector and a FaceEncoder model is required to use this function.
warmUp(int imageWidth, int imageHeight) → void
Gets the face tracker module ready to work on a specific image size by initializing all its internal memory layout. This function can be called after setting all the parameters of the face detector module if the image size is fixed, for example using frames from a video stream. Calling this function is not mandatory as the internal memory layout is automatically initialized in the first call to detectFaces() or trackFaces() if missing. Important: Loading a FaceDetector model is required to use this function.

Operators

operator ==(Object other) → bool
The equality operator.
inherited