I have a lot of binary images to OCR. They were created by binarizing bitmaps where the background was black and text white, but the binarizer was tuned for black text and a white background.
For the vast majority of these, I have no access to the original items or the original bitmaps.
Unfortunately, the binarization process created a binary image with a residual shadow or halo around all of the previously white text. I think that this is common artifact of white text on black background image binarization and it is an impediment to reading the text using standard OCR techniques.
I want to OCR this kind of shadowed text using OpenCV, but have been unable to find the appropriate terminology to search out solutions, if any, from others with a similar issue.
Does this affect have a common name in the OCR world?
Here is an example of what I am seeing: Well I cannot yet post photos... Too new :-)
somewhat related, but not identical to the issue that I describe:
How to remove Blur halo effect
Sharpening image using OpenCV OCR
