Tesseract OCR ANE

Tesseract OCR Adobe Native Extension for Android & iOS.

Getting Started

NB: You'll need trained data files (https://github.com/tesseract-ocr/tessdata) to be located in tessdata directory (deployed together with your project or already available on target device) in order Tesseract engine to work.

In order to use Tesseract OCR in your AS3/AIR project:

1. Add ANE to your project.

2. Instatntiate the main ANE class:

   _ane = TesseractANE.getInstance();
   

3. Setup recognition params:

   • (required!) absolute path to /tessdata folder, meaning "/storage/sdcard0/data/" to be provided if your tess data files are located at /storage/sdcard0/data/tessdata/ folder:

   _ane.absoluteTessdataPath = "<path>"; 
   

   • (optional, "eng" by default) set of languages to be used for recognition:

   _ane.language = "eng+rus";
   

   • (optional, "1234567890ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz" by default) set of chars to be recognized:

   _ane.charset = "123ABC";
   

4. Setup recognition result listener:

   _ane.addEventListener(TesseractANEEvent.RECOGNIZED, onTextRecognized);
   ...
   protected function onTextRecognized(event:TesseractANEEvent):void {
       trace("end recoignizing " + event.uuid + " > " + event.text);
   }
   

5. Pass BitmapData instance for OCR processing:

   var uuid : String = _ane.recognize(bitmap.bitmapData));
   

   The method will return UUID string value that can be used for identification in case you start some recognition processes simultaneousely.

6. After the recognition process completes you'll receive TesseractANEEvent.RECOGNIZED event, containing recognized text and corresponding process UUID value.

   trace("end recoignizing " + event.uuid + " > " + event.text);
   

Usage Tips

• The whole available set of trained data files require a lot of disk space, it's better to deploy files only for languages that you are going to use (see file names).

Customization Tips

• Update path variables at the beginning of buildANE script for your platform (.bat for Win, .sh for OSX) before trying to rebuild ANE.
• All the build scripts are processing files only inside the build folder and subfolders, so you'll need to copy all updated *.a, *.so or *.aar there yourself after making changes in native projects.
• Due to the big size of resulting ANE file you can change buildANE script for your platform (.bat for Win, .sh for OSX) and extension.xml and rebuild ANE just for Android or iOS.

License

Tesseract OCR Adobe Native Extension is distributed under the Apache 2.0 license (see LICENSE.md).

Tesseract OCR iOS and TesseractOCR.framework, maintained by Daniele Galiotto (founder) (http://www.g8production.com), are distributed under the MIT license (see https://opensource.org/licenses/MIT).

Tess-two, maintained by Robert Theis (support page https://stackoverflow.com/questions/tagged/tess-two), is distributed under the Apache 2.0 license (see http://www.apache.org/licenses/LICENSE-2.0).

Tesseract, maintained by Google (http://code.google.com/p/tesseract-ocr/), is distributed under the Apache 2.0 license (see http://www.apache.org/licenses/LICENSE-2.0).