android - Optimize an image for text recognition using tesseract -
i have user tesseract ocr android project recognize text image taken camera. results not accurate. want optimize image using opencv. want achieve following captured image decoded in bitmap.config.argb_8888 format:
- detect objects in resized image.
- once object identified, compute border w.r.t original image. (this removing camera angle effect)
- extract object original image, applying perspective transform.
- apply white balance remove lightening effects.
in example provided tess_two api, using leptonica image manipulations drawing bounding boxes around words..but in case want use opencv...your guidance highly appreciated...
that's lot asking for, , depending on object may impossible. should check out tutorials on 2d feature detection , object detection (http://docs.opencv.org/doc/tutorials/features2d/table_of_content_features2d/table_of_content_features2d.html , http://docs.opencv.org/doc/tutorials/objdetect/table_of_content_objdetect/table_of_content_objdetect.html) see if there can use.
white balance not lighting, should adaptive thresholding or kind of high pass filtering instead.
Comments
Post a Comment