tzutalin / dlib-android-app

dlib-android-app

See http://dlib.net for the main project documentation.

See dlib-android for JNI lib. Refer to dlib-android/jni/jnilib_ex

Grap the source

$ git clone https://github.com/tzutalin/dlib-android-app.git

Features

• Support HOG detector

• HOG Face detection

• Facial Landmark/Expression

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

Update shared lib (Optional)

You can build shared library from dlib-android

Copy the shared library to ./dlib/src/main/jniLibs/

Try directly

Download and 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 {
    implementation 'com.tzutalin.dlib-android-app:dlib:1.0.4'
}

Sample code

Facial landmark detection

FaceDet faceDet = new FaceDet(Constants.getFaceShapeModelPath());
Bitmap bitmap = BitmapFactory.decodeFile("Image Path");
List<VisionDetRet> results = faceDet.detect(bitmap);
for (final VisionDetRet ret : results) {
    String label = ret.getLabel();
    int rectLeft = ret.getLeft();
    int rectTop = ret.getTop();
    int rectRight = ret.getRight();
    int rectBottom = ret.getBottom();
    // Get 68 landmark points
    ArrayList<Point> landmarks = ret.getFaceLandmarks();
    for (Point point : landmarks) {
        int pointX = point.x;
        int pointY = point.y;
    }
}

Pedestrian detection

Pedestrian pedestrianDet = new PedestrianDet();
List<VisionDetRet> personList = pedestrianDet.detect(imgPath);

License

License