Hey.
as far as I understand tesseract acts on grayscale images with one
byte per pixel. This works for me (given an rgb image):
cv::cvtColor( image, gray, CV_RGB2GRAY );
tess_api.TesseractRect( gray.data, 1, grey.step1(), 0, 0,
gray.cols, gray.rows );
const char * txt = tess_api.GetUTF8Text();
the output may well depend on how much stuff there is in the image
besides the text, since tesseract seems to apply a single threshold to
the whole image, and if your writing and its background are not
separated this threshold you are out of luck. It should work if the
image is covered with the text page, consisting of essentially two
brightness levels.
On 14 Nov., 19:33, cyrt <[email protected]> wrote:
> I'd like to perform OCR on subimages that I am loading using OpenCV.
> How can I convert this data into a form tesseract can work with?
>
> So far I tried two different codes, but no version is working. I hope
> someone here can help me finding a solution.
>
> A.) converting cv::Mat into Pix*
>
> cv::Mat image = cv::imread("c:/image.png");
> cv::Mat subImage = image(cv::Rect(50, 200, 300, 100));
>
> int depth;
> if(subImage.depth() == CV_8U)
> depth = 8;
> //other cases not considered yet
>
> PIX* pix = pixCreateHeader(subImage.size().width,
> subImage.size().height, depth);
> pix->data = (l_uint32*) subImage.data;
>
> tesseract::TessBaseAPI tess;
> STRING text;
> if(tess.ProcessPage(pix, 0, 0, &text))
> {
> std::cout << text.string();
> }
>
> OCR returns only non-readable characters however.
>
> B.) Using cv::Mat for TesseractRect()
>
> cv::Mat image = cv::imread("c:/image.png");
> cv::Mat subImage = image(cv::Rect(50, 200, 300, 100));
> char* cr = tess.TesseractRect(
> subImage.data,
> subImage.channels(),
> subImage.channels() * subImage.size().width,
> 0,
> 0,
> subImage.size().width,
> subImage.size().height);
>
> This code doesn't either and also returns only non-readable
> characters, although different ones than from the code above.
>
> Does anyone know what the problem could be? cv::Mat stores pixel data
> as an array of type uchar, so it should be fine to use in
> TesseractRect without any conversion, as UINT8* are required.
--
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en
