Reputation: 847
I'm using ml kit cloud text recognition by java, and it works perfectly for all languages except Gujarati.
i cant understand whats wrong, i did also add "gu" language to recognition options but it didn't matter. whats wrong?
FirebaseVisionImage visionImage = FirebaseVisionImage.fromBitmap(myBitmap);
FirebaseVisionCloudTextRecognizerOptions options = new FirebaseVisionCloudTextRecognizerOptions.Builder()
.setLanguageHints(Arrays.asList("gu"))
.build();
FirebaseVisionTextRecognizer detector = FirebaseVision.getInstance()
.getCloudTextRecognizer(options);
Task<FirebaseVisionText> result =
detector.processImage(visionImage)
.addOnSuccessListener(new OnSuccessListener<FirebaseVisionText>() {
@Override
public void onSuccess(FirebaseVisionText firebaseVisionText) {
Log.e("Recognition", "Text : " + firebaseVisionText.getText());
}
})
.addOnFailureListener(
new OnFailureListener() {
@Override
public void onFailure(@NonNull Exception e) {
Log.e(TAG, "Recognition failed : " + e.getMessage());
}
});
Upvotes: 1
Views: 796
Reputation: 935
Have you tried the SPARSE_MODEL without the language hint? It should automatically detect the language. There is a known internal issue with 'gu' hint for SPARSE_MODEL, and we are working on it.
Also, you could also try to use DENSE_MODEL instead of SPARSE_MODEL with the language hint.
FirebaseVisionCloudTextRecognizerOptions options = new FirebaseVisionCloudTextRecognizerOptions.Builder()
.setLanguageHints(Arrays.asList("gu"))
.setModelType(FirebaseVisionCloudTextRecognizerOptions.DENSE_MODEL)
.build();
Upvotes: 1
Reputation: 847
I had communications with cloud support, and it turned out that problem is from their side, and they are working on that.
Upvotes: 1