[Android] OCR - 使用 Tesseract

[參考資料]

http://honglung.pixnet.net/blog/post/165953088-android-%E5%88%A9%E7%94%A8-tesseract-ocr-%E9%80%B2%E8%A1%8C-%E6%96%87%E5%AD%97%E8%BE%A8%E8%AD%98

https://github.com/tesseract-ocr/tesseract

[需要權限]

<uses-permission android:name="android.permission.WRITE_EXTERNAL_STORAGE"/>

[Gradle]
build.gradle (Module:app)

implementation 'com.rmtheis:tess-two:9.0.0'

[使用方式]

public String runOcr( Bitmap bmp){

    final TessBaseAPI ocrApi = new TessBaseAPI();
    ocrApi.init( pathTessBase, "chi_tra");
    ocrApi.setPageSegMode(TessBaseAPI.PageSegMode.PSM_AUTO);
    ocrApi.setImage(bmp);
    String resString = ocrApi.getUTF8Text();

    ocrApi.clear();
    ocrApi.end();
    return  resString;
}

[註]
1. 訓練資料檔 chi_tra.traineddata 必須放在 pathThessBase/tessdata 資料夾下.

2. PageSegmentationMode = TessBaseAPI.PageSegMode.PSM_SINGLE_LINE 只可辨識單行文字

    PageSegmentationMode = TessBaseAPI.PageSegMode.PSM_AUTO 可辨識多行文字