HI, All
Please, suggest who is an expert in graphical algorithms: how should image be pre-processed to make OCR better ?
Common algorithm.
I guess,
1) Convert an image to greyscale one
2) Maybe to check in whole if the background is light, or dark. If dark background - to invert whole image, to get text of dark color.
3) And ... what to next before recognition try ?
Some filtration, brightness, contrast changes ?
Please, suggest who is an expert in graphical algorithms: how should image be pre-processed to make OCR better ?
Common algorithm.
I guess,
1) Convert an image to greyscale one
2) Maybe to check in whole if the background is light, or dark. If dark background - to invert whole image, to get text of dark color.
3) And ... what to next before recognition try ?
Some filtration, brightness, contrast changes ?
Last edited: