Firebase ML 套件无法准确识别语言（英语或其他）

Question

I am developing a text recognizer app for android using firebase ML Kit. The issue when I capture the text image is not getting the actual result of the image why??. I think my app does not identify the language well. How can I solve this issue?

#I am new to android Pls help me in this regard

Example

Dependency I have used

implementation 'com.google.firebase:firebase-ml-vision:20.0.0'

Recognize Text & Process Text code

/**
 * This method is used to extract the text from the image
 * @param image
 */
private void recognizeText(FirebaseVisionImage image){
    FirebaseVisionTextRecognizer detector = FirebaseVision.getInstance()
            .getOnDeviceTextRecognizer();
    detector.processImage(image)
            .addOnSuccessListener(new OnSuccessListener<FirebaseVisionText>() {
                @Override
                public void onSuccess(FirebaseVisionText firebaseVisionText) {
                    processText(firebaseVisionText);
                }
            })
            .addOnFailureListener(new OnFailureListener() {
                @Override
                public void onFailure(@NonNull Exception e) {
                    e.printStackTrace();
                    Toast.makeText(getActivity(), "No text found", Toast.LENGTH_SHORT).show();
                }
            });
}

/**
 * This method is used to process the text and send the result (text) to ResultLayout fragment
 * @param firebaseVisionText
 */
private void processText(FirebaseVisionText firebaseVisionText) {
    if(firebaseVisionText.getTextBlocks().isEmpty()){
        Toast.makeText(getActivity(), "No text found or Text may not be clear", Toast.LENGTH_LONG).show();
    }
    else {
        String text = "";
        for (FirebaseVisionText.TextBlock block : firebaseVisionText.getTextBlocks()){
            text = text + block.getText() + " ";
        }

        String tag = ((MainActivity)getActivity()).getTag_ResultLayout();
        ResultLayout fragment = (ResultLayout)getActivity().getSupportFragmentManager()
                .findFragmentByTag(tag);
        fragment.setResult(text);
        ((MainActivity)getActivity()).openResultLayout();
    }

I am ready to provide more information in this regard

Answer 1

Could you share your AndroidManifest.xml?

If you want to achieve better results, you can try to use multiple models like this:

  <application ...>
      ...
      <meta-data
          android:name="com.google.firebase.ml.vision.DEPENDENCIES"
          android:value="ocr" />
      <!-- To use multiple models: android:value="ocr,model2,model3" -->
  </application>

More info on this can be found in the official documentation here: https://firebase.google.com/docs/ml-kit/android/recognize-text

Answer 2

we made some changes to Firebase ML Kit for Firebase to better distinguish the on-device APIs from cloud based APIs. "ML Kit"(without firebase branding), contains all the on-device APIs. Here's the migration guide from firebase mlkit to mlkit.

This issue might be resulted in a misformatted image. Could you try to refer to the sample here for constructing a InputImage from different sources?