Removing caption background for use in ORC (tessearct)

[gilberto_san]

Try this Unix syntax:

Code: Select all

convert 9a6c443dfa76445fa95158948cc4e9bd.png \
\( -clone 0 -colorspace HCL -channel g -separate +channel -threshold 40% \) \
-alpha off -compose copy_opacity -composite result.png

Please always provide your IM version and platform, since syntax may differ.

I´m testing
tks

[gilberto_san]

I used part of your command

Code: Select all

convert print1.bmp -colorspace HCL -channel g -separate +channel -threshold 10%% -negate print1.bmp

And got 3 figures. (print1-0.bmp print1-1.bmp print1-2.bmp)
Figure print1-1.bmp is perfect for my purpose.

Now 90% of my problem is solved. thank you so much.

Let's go to the remaining 10%.
I have images with the gray background and the font with shades of gray.
Serpá that I can remove the bottom?
Remembering that I can make multiple prints with the same text. And different backgrounds
Do you have any tips or suggestions for such a problem?
Thanks

[fmw42]

I do not understand your question. Please post examples and explain further.

[gilberto_san]

i use the command

Code: Select all

 convert aaaa.bmp -colorspace HCL -channel g -separate +channel -threshold 10% img.bmp  

And got 3 pictures : img-0.bmp | img-1.bmp |img-2.bmp.

Figure img-1.bmp is great for character recognition

If I use in a gray font the result is bad


Any suggestion to be able to extract the character when the font is gray?

[fmw42]

Gray images have no saturation (since there is no color), so the same technique will not help. All you can do it threshold on intensity.

Code: Select all

convert image -threshold XX% result

[fmw42]

If on Unix system, you could try some of my auto-thresholding techniques. See my home page in the table of categories down the page.