Flutter using Google ML-Kit

Using Google’s ML-Kit is really simple.

First of all add ML-Kit library into pubspec.yaml. Besides, to be able to use camera easily, I added camera_camera library, too.

google_ml_kit: ^0.16.2
camera_camera: ^3.0.0

Barcode & QR Code

To read barcodes and QR codes, we need to use BarcodeScanner.

BarcodeScanner barcodeScanner = BarcodeScanner();

After the above declaration, now, we can call the method to read barcode. Be aware that, the method is to be async.

/// Scan barcode or QR code
Future<void> scanBarcode() async {
  List<String?> barcodes = [];

  InputImage image = InputImage.fromFile(imageFile);

  List<Barcode> barcodeList = await barcodeScanner.processImage(image);

  if (barcodeList.isNotEmpty) {
    for (Barcode barcodeLabel in barcodeList) {
      barcodes.add(barcodeLabel.rawValue);
    }
  }
}

Text Recognition

Text recognition will yield the ability of parsing texts just from images.

To start, let us declare our variable.

TextRecognizer recognizer = TextRecognizer();

Now, call the necessary method to start reading texts.

/// Detect texts
Future<void> fetchText() async {
  List<String> parsedTextList = [];

  InputImage image = InputImage.fromFile(imageFile);

  RecognizedText recognisedText = await recognizer.processImage(image);

  for (TextBlock block in recognisedText.blocks) {
    for (TextLine line in block.lines) {
      parsedTextList.add(line.text);
    }
  }
}

Image Labeling

Image labeling reads the features of an image.

To declare image labeler, first, we need to declare image labeler options.

final ImageLabelerOptions options = ImageLabelerOptions(confidenceThreshold: 0.5);
late ImageLabeler imageLabeler;

Here, we gave a confidence threshold of %50. This means, the detected features with minimum %50 probability will be retrieved from the library. Features that are detected with lower probability (confidence) will not be listed.

/// Read labels
Future<void> readImageLabels() async {
  List<String> imageLabels = [];

  InputImage image = InputImage.fromFile(imageFile);

  imageLabeler = ImageLabeler(options: options);
  List<ImageLabel> labels = await imageLabeler.processImage(image);

  for (ImageLabel label in labels) {
    imageLabels.add(label.label);
    print("Feature: ${label.label} with confidence of ${label.confidence}");
  }
}

That is all.

Please do not forget to close ML-Kit variables during dispose.

@override
void dispose() {
  barcodeScanner.close();
  recognizer.close();
  imageLabeler.close();

  super.dispose();
}

By the way, let us see the easy usage of camera_camera, too.

void openCamera() {
  Navigator.push(
      context,
      MaterialPageRoute(
          builder: (_) => CameraCamera(
            onFile: (file) {
              imageFile = file;
              Navigator.pop(context);
              // todo call your ml-kit method here
            },
          )));
}

That is all for calling camera retrieving the photo file which was taken with the device’s camera.