eu.id3.face

Class FacePad

java.lang.Object

eu.id3.face.FacePad

All Implemented Interfaces:

java.io.Serializable, java.lang.AutoCloseable

public class FacePad
extends java.lang.Object
implements java.lang.AutoCloseable, java.io.Serializable

Detects presentation attacks in still images or video frames.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
FacePad() Creates a new instance of the FacePad class.
FacePad(long handle) Creates a new instance of the FacePad class.

Method Summary

Modifier and Type	Method and Description
void	close() Releases all resources used by this FacePad.
int	computeBlurrinessScore(Image image, DetectedFace detectedFace) Computes a blurriness score for a color image.
ColorBasedPadResult	computeColorBasedScore(Image image, DetectedFace detectedFace) Computes the PAD score and confidence of a detected face using only the color image.
int	computeDepthBasedScore(Image image, DetectedFace detectedFace) Computes the PAD score of a detected face using a depth map image.
DetectedFaceAttackSupport	detectAttackSupport(Image image, DetectedFace detectedFace) Detects if an attack support surrounds the detected face.
protected void	finalize() Finalizer for this class.
int	getAttackSupportDetectorConfidenceThreshold() Gets the attack support detector confidence threshold, in the range [0;100].
long	getHandle() Gets the native handle.
void	setAttackSupportDetectorConfidenceThreshold(int attackSupportDetectorConfidenceThreshold) Sets the attack support detector confidence threshold, in the range [0;100].

Methods inherited from class java.lang.Object

clone, equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

FacePad

public FacePad()
        throws FaceException

Creates a new instance of the FacePad class.

Throws:

FaceException - An error has occurred during Face Library execution.

FacePad

public FacePad(long handle)
        throws FaceException

Creates a new instance of the FacePad class.

Parameters:

handle - Handle to the FacePad.

Throws:

FaceException - An error has occurred during Face Library execution.

Method Detail

close

public void close()

Releases all resources used by this FacePad.

Specified by:

close in interface java.lang.AutoCloseable

finalize

protected void finalize()

Finalizer for this class. Finalizers are deprecated since Java 9+. This method is not guaranteed to be called. To release all memory held by this class please be sure to call the close() method.

Overrides:

finalize in class java.lang.Object

getHandle

public long getHandle()

Gets the native handle.

Returns:

The native handle.

getAttackSupportDetectorConfidenceThreshold

public int getAttackSupportDetectorConfidenceThreshold()
                                                throws FaceException

Gets the attack support detector confidence threshold, in the range [0;100]. Hint: Default value is 25. Note: Setting a high threshold reduces false attack support detections but can increase the number of undetected attack supports.

Returns:

Attack support detector confidence threshold, in the range [0;100].

Throws:

FaceException - An error has occurred during Face Library execution.

setAttackSupportDetectorConfidenceThreshold

public void setAttackSupportDetectorConfidenceThreshold(int attackSupportDetectorConfidenceThreshold)
                                                 throws FaceException

Sets the attack support detector confidence threshold, in the range [0;100]. Hint: Default value is 25. Note: Setting a high threshold reduces false attack support detections but can increase the number of undetected attack supports.

Parameters:

attackSupportDetectorConfidenceThreshold - Attack support detector confidence threshold, in the range [0;100].

Throws:

FaceException - An error has occurred during Face Library execution.

computeBlurrinessScore

public int computeBlurrinessScore(Image image,
                                  DetectedFace detectedFace)
                           throws FaceException

Computes a blurriness score for a color image. An attack presented on a low resolution support has more chance to be blurred than a bonafide. The maximum recommended value is 20. Important: Loading the model ``FaceBlurrinessDetector`` is required to use this function.

Parameters:

image - Source image to process.

detectedFace - Detected face to process.

Returns:

The computed blurriness score, in the range [0;100].

Throws:

FaceException - An error has occurred during Face Library execution.

computeColorBasedScore

public ColorBasedPadResult computeColorBasedScore(Image image,
                                                  DetectedFace detectedFace)
                                           throws FaceException

Computes the PAD score and confidence of a detected face using only the color image. A high score means a high probability for the face to be bona-fide, hence not an attack. The minimum recommended score is 90. A low confidence means that the quality of the image is not sufficient enough to take a decision. The minimum recommended confidence is 70. Warning: A minimum IOD (64 pixels per default) for the detected face is required for this function, below this value it will output an error. Important: Loading the ``FaceColorBasedPad`` model is required to use this function.

Parameters:

image - Source image to process.

detectedFace - Detected face to process.

Returns:

The computed PAD result including score and confidence.

Throws:

FaceException - An error has occurred during Face Library execution.

computeDepthBasedScore

public int computeDepthBasedScore(Image image,
                                  DetectedFace detectedFace)
                           throws FaceException

Computes the PAD score of a detected face using a depth map image. A high score means a high probability for the face to be bona-fide, hence not an attack. The minimum recommended value is 10. Important: Loading the ``FaceDepthBasedPad`` model is required to use this function.

Parameters:

image - Source image to process. Must be Grayscale 16 Bits.

detectedFace - Detected face to process.

Returns:

The computed PAD score, in the range [0;100].

Throws:

FaceException - An error has occurred during Face Library execution.

detectAttackSupport

public DetectedFaceAttackSupport detectAttackSupport(Image image,
                                                     DetectedFace detectedFace)
                                              throws FaceException

Detects if an attack support surrounds the detected face. The recommended usage is to consider the image to be an attack as soon as an attack support is detected. Important: Loading the ``FaceAttackSupportDetector`` model is required to use this function.

Parameters:

image - Source image to process.

detectedFace - Detected face to process.

Returns:

The detected face attack support.

Throws:

FaceException - An error has occurred during Face Library execution.