Examples with MultiProcessor com.google.android.gms.vision.MultiProcessor used on opensource projects

Example 1 with MultiProcessor

use of com.google.android.gms.vision.MultiProcessor in project android-vision by googlesamples.

the class GooglyEyesActivity method createFaceDetector.

//==============================================================================================
// Detector
//==============================================================================================
/**
     * Creates the face detector and associated processing pipeline to support either front facing
     * mode or rear facing mode.  Checks if the detector is ready to use, and displays a low storage
     * warning if it was not possible to download the face library.
     */
@NonNull
private FaceDetector createFaceDetector(Context context) {
    // For both front facing and rear facing modes, the detector is initialized to do landmark
    // detection (to find the eyes), classification (to determine if the eyes are open), and
    // tracking.
    //
    // Use of "fast mode" enables faster detection for frontward faces, at the expense of not
    // attempting to detect faces at more varied angles (e.g., faces in profile).  Therefore,
    // faces that are turned too far won't be detected under fast mode.
    //
    // For front facing mode only, the detector will use the "prominent face only" setting,
    // which is optimized for tracking a single relatively large face.  This setting allows the
    // detector to take some shortcuts to make tracking faster, at the expense of not being able
    // to track multiple faces.
    //
    // Setting the minimum face size not only controls how large faces must be in order to be
    // detected, it also affects performance.  Since it takes longer to scan for smaller faces,
    // we increase the minimum face size for the rear facing mode a little bit in order to make
    // tracking faster (at the expense of missing smaller faces).  But this optimization is less
    // important for the front facing case, because when "prominent face only" is enabled, the
    // detector stops scanning for faces after it has found the first (large) face.
    FaceDetector detector = new FaceDetector.Builder(context).setLandmarkType(FaceDetector.ALL_LANDMARKS).setClassificationType(FaceDetector.ALL_CLASSIFICATIONS).setTrackingEnabled(true).setMode(FaceDetector.FAST_MODE).setProminentFaceOnly(mIsFrontFacing).setMinFaceSize(mIsFrontFacing ? 0.35f : 0.15f).build();
    Detector.Processor<Face> processor;
    if (mIsFrontFacing) {
        // For front facing mode, a single tracker instance is used with an associated focusing
        // processor.  This configuration allows the face detector to take some shortcuts to
        // speed up detection, in that it can quit after finding a single face and can assume
        // that the nextIrisPosition face position is usually relatively close to the last seen
        // face position.
        Tracker<Face> tracker = new GooglyFaceTracker(mGraphicOverlay);
        processor = new LargestFaceFocusingProcessor.Builder(detector, tracker).build();
    } else {
        // For rear facing mode, a factory is used to create per-face tracker instances.  A
        // tracker is created for each face and is maintained as long as the same face is
        // visible, enabling per-face state to be maintained over time.  This is used to store
        // the iris position and velocity for each face independently, simulating the motion of
        // the eyes of any number of faces over time.
        //
        // Both the front facing mode and the rear facing mode use the same tracker
        // implementation, avoiding the need for any additional code.  The only difference
        // between these cases is the choice of Processor: one that is specialized for tracking
        // a single face or one that can handle multiple faces.  Here, we use MultiProcessor,
        // which is a standard component of the mobile vision API for managing multiple items.
        MultiProcessor.Factory<Face> factory = new MultiProcessor.Factory<Face>() {

            @Override
            public Tracker<Face> create(Face face) {
                return new GooglyFaceTracker(mGraphicOverlay);
            }
        };
        processor = new MultiProcessor.Builder<>(factory).build();
    }
    detector.setProcessor(processor);
    if (!detector.isOperational()) {
        // Note: The first time that an app using face API is installed on a device, GMS will
        // download a native library to the device in order to do detection.  Usually this
        // completes before the app is run for the first time.  But if that download has not yet
        // completed, then the above call will not detect any faces.
        //
        // isOperational() can be used to check if the required native library is currently
        // available.  The detector will automatically become operational once the library
        // download completes on device.
        Log.w(TAG, "Face detector dependencies are not yet available.");
        // Check for low storage.  If there is low storage, the native library will not be
        // downloaded, so detection will not become operational.
        IntentFilter lowStorageFilter = new IntentFilter(Intent.ACTION_DEVICE_STORAGE_LOW);
        boolean hasLowStorage = registerReceiver(null, lowStorageFilter) != null;
        if (hasLowStorage) {
            Toast.makeText(this, R.string.low_storage_error, Toast.LENGTH_LONG).show();
            Log.w(TAG, getString(R.string.low_storage_error));
        }
    }
    return detector;
}