Theo's Typesafe Cult•12mo ago

OCR (Tesseract) not working well with very clear text (in c or js)

Wondering if anyone has any experience with OCR. I'm trying to get tesseract OCR to read the names of the characters off of this image. I am pre-processing the images in an effort to make it easier (see the other 2 images), but still haven't had any luck. I am currently using the following settings:

    TessBaseAPISetPageSegMode(autovod->tess, PSM_SINGLE_BLOCK);
    TessBaseAPISetSourceResolution(autovod->tess, 700);
    TessBaseAPISetVariable(autovod->tess, "language_model_penalty_non_dict_word", "0");
    TessBaseAPISetVariable(autovod->tess, "tessedit_char_whitelist", "ABCDEFGHIJKLMNOPQRSTUVWXYZ");

    TessBaseAPISetPageSegMode(autovod->tess, PSM_SINGLE_BLOCK);
    TessBaseAPISetSourceResolution(autovod->tess, 700);
    TessBaseAPISetVariable(autovod->tess, "language_model_penalty_non_dict_word", "0");
    TessBaseAPISetVariable(autovod->tess, "tessedit_char_whitelist", "ABCDEFGHIJKLMNOPQRSTUVWXYZ");

Anybody have any idea what could be going wrong?

0 Replies

No replies yetBe the first to reply to this messageJoin

Gaming

Programming

OCR (Tesseract) not working well with very clear text (in c or js)