permission to use the model in an app

[talutalu]

Hi CaptainDario,

Maybe it's a dumb question but I want to use your tansorflow lite model in my japanese dictionary app and I'm asking your permission to do so.

I think that your solution to recognise handwritten kanjis is better done than in most of the dictionary apps I know. I know that you already released an android app to recognise a single kanji so I wasn't sure if I could integrate the model into mine.

Thank you for your awesome work!
Christophe

[CaptainDario]

Hi talutalu,

thank you for the kind words.

Of course you are welcome to use this model in your app.
In the end I open sourced this model because I want it to be used from more people.
However I would only ask you to give me credit somwhere in your app with smth like:
Character recognition powered by machine learning from CaptainDario (DaAppLab)
I also added a new section to the README on how to include this in your own software.

I also added your app to the table of apps which use this model (however if this bothers you I can remove it).

[talutalu]

Thank you!

I will add that line in the about section of my app when the integration is done!

I'm far from being an expert of tensorflow so I struggled a bit to integrate the model into my code. It seems to work now so I thought that this piece of code could help others to integrate the model in their android app:

class HandWritingService {

    val interpreter: Interpreter;
    val labels: List<Char>;

    init {
        interpreter = Interpreter(File(App.instance.handWritingModelLocation));
        labels = File(App.instance.handWritingLabelsLocation).readText().chars().mapToObj {
            Char(it)
        }.collect(Collectors.toList());
    }

    fun recognize(bitmap: Bitmap): List<Pair<Char, Float>> {
        val size = bitmap.width;
        val inputBuffer = TensorBuffer.createFixedSize(intArrayOf(1, size, size, 1), DataType.FLOAT32);
        val intArray = IntArray(size * size);
        bitmap.getPixels(intArray, 0, size, 0, 0, size, size);
        val floatArray = FloatArray(size * size);
        for (i in intArray.indices) {
            // we select the alpha channel of the pixel
            floatArray[i] = ((intArray[i] shr 24) and 0x000000FF).toFloat();
        }
        inputBuffer.loadArray(floatArray);
        interpreter.resizeInput(0, intArrayOf(1, size, size, 1));

        val numOutputs = interpreter.getOutputTensor(0).numElements();
        val probabilityBuffer = TensorBuffer.createFixedSize(intArrayOf(1, numOutputs), DataType.FLOAT32);

        interpreter.run(inputBuffer.buffer, probabilityBuffer.buffer);
        val result = probabilityBuffer.floatArray;

        return result.mapIndexed { i, v -> Pair(labels[i], v) }.sortedByDescending { it.second };
    }
}

[talutalu]

Hi CaptainDario,

The update of my app is now live. Thank you again for your work!

I tried to make some funny example to show that the stroke order or count is not that important:

[CaptainDario]

Thank zou for the kind words and congrats on the nice app!
I hope you will get a lot of users!