gaze recognition

• Stage 1:Facial Landmark Detection
• Stage 2:Gaze tracking within an error range of 2 centimeters (so only applicable on Tablet so far)

Features

• Support HOG detector on Facial Landmark Detection
• Use linear suporposition of images to track eyesight

Demo

Build

Android app

• Open Android studio to build

• Use command line to build (Optional)

On Windows platforms, type this command:

$ gradlew.bat assembleDebug

On Mac OS and Linux platforms, type these commands:

$ ./gradlew assembleDebug

or

$ make ; make install

Try directly

Install the apk

$ adb install demo/app-debug.apk

Otherwise, import the library to your build.gradle

repositories {
    maven {
        url 'https://dl.bintray.com/tzutalin/maven'
    }
}

dependencies {
    compile 'com.tzutalin.dlib-android-app:dlib:1.0.3'
}

Sample code

Facial landmark detection

PeopleDet peopleDet = new PeopleDet();
List<VisionDetRet> results = peopleDet.detBitmapFace(bitmap, Constants.getFaceShapeModelPath());
for (final VisionDetRet ret : results) {
    String label = ret.getLabel(); // If doing face detection, it will be 'Face'
    int rectLeft = ret.getLeft();
    int rectTop= ret.getTop();
    int rectRight = ret.getRight();
    int rectBottom = ret.getBottom();
    ArrayList<Point> landmarks = ret.getFaceLandmarks();
    for (Point point : landmarks) {
        int pointX = (int) (point.x * resizeRatio);
        int pointY = (int) (point.y * resizeRatio);
        // Get the point of the face landmarks
    }
}

License

License