Android

Technical Specifications:

You can implement DoreEmotionRecognition SDK in your existing/new Android project with in just 3 Steps

Step 1:

 1. Add this URL in project -> build.gradle  (under respositories section)

allprojects {
    repositories {
        google()
        jcenter()
        maven { url "https://raw.github.com/doreai/Framework/master" }
    }
}

2. Add this line in app -> build.gradle  (under dependencies section) 

   Version number will be vary depends on latest release 

dependencies {
    ....
     
    implementation 'com.dore:emotionrecognition:1.0.1'

    
}

 Don't forgot to Sync your project 

Step 2: 

Initiate DoreEmotionRecognition Manager

Kotlin Code :

private val bEngine: EmotionRecognitionManager = EmotionRecognitionManager()

val lickeycode = "insert your license key here"
bEngine.init_data(this,lickeycode)

Java Code :

EmotionRecognitionManager bEngine = new EmotionRecognitionManager();

String lickeycode = "insert your license key here";
bEngine.init_data(this,lickeycode);

Step 3:

Execute EmotionRecognition Manager with input detected Face

Process 1 :  Face detection with bounding box, crop image with bounding box value, and convert to grayscale image

Process 2 :  Emotion Recognition with input cropped image

Use TextureView.SurfaceTextureListener (onSurfaceTextureUpdated) to get camera frame image

Kotlin Code : 

 override fun onSurfaceTextureUpdated(surfaceTexture: SurfaceTexture) {
        val rlist = bEngine.detectFace(tex_view.bitmap)

        if(rlist.count() > 0){
            val croppedBmp = bEngine.cropImage(tex_view.bitmap,rlist[0])
            val bwImg = bEngine.convertToGrayScale(croppedBmp)
            val dimage = DoreImage.fromBitmap(bwImg)
            var result = bEngine.run(dimage)

            Log.d("emotion", result?.getClsName())

            runOnUiThread {
                outImg.setImageBitmap(croppedBmp)
                lbResultTxt.text = result?.getClsName()
            }
        }else{
            runOnUiThread{
                lbResultTxt.text = "Face Not Found"
            }
        }
 }

Java Code :

 public void onSurfaceTextureUpdated(SurfaceTexture surface) {
          List<Rect> rlist = bEngine.detectFace(tex_view.bitmap);
          if(rlist.size() > 0){
             Bitmap croppedBmp = bEngine.cropImage(tex_view.bitmap,rlist[0]);
             Bitmap bwImg = bEngine.convertToGrayScale(croppedBmp);
             DoreImage dimage = DoreImage.fromBitmap(bwImg);
             EmotionResult result = bEngine.run(dimage);
             Log.d("emotion :" ,result.getClsName());
              runOnUiThread(new Runnable() {
                  @Override
                  public void run() {
                      outImg.setImageBitmap(croppedBmp);
                      lbResultTxt.text =  result.getClsName();
                  }
              });


          }else {
              runOnUiThread(new Runnable() {
                  @Override
                  public void run() {
                      lbResultTxt.text = "Face Not Found";
                  }
              });

          }

    }

Emotion Text Result values - "Happy", "Angry", "Surprise", "Sad", "Neutral"

Additional Note: Supported ABIs :  armeabi-v7a and arm64-v8a, which should cover most modern Android devices.