Is there an option to remove shadow from scanned documents

[Sicos1977]

Since I'm a newbie with Magick.NET I have this question. Is there an option to remove shadow from a scanned document.

See the image below (I cant show the complete image because there is personal information on it). When I use OTSU thresholding on this then to lower shadow part turns black so I was wondering if there is an option to remove the shadow before using OTSU?

[dlemstra]

You probably want to ask that question here instead: https://github.com/ImageMagick/ImageMagick/discussions. I can then help you at a later moment to translate those commands into Magick.NET method calls if you need some help with.

[Sicos1977]

Going to do that, thanks.

It seems a very difficult problem to fix, also tested this with Tesseract OCR own image optimalisation and that also turned it to a block of black.

I use ImageMagick.Net to cleanup declarations that people sent in that they have photographed with their smartphone. I understand there is no 100% solution for everything because they sent in to much junk :-)

[Sicos1977]

        private void RemoveShadow(MagickImage image)
        {
            // convert 131828504-61b59e6a-25f8-405c-9366-6158baeccabe.png -negate -lat 20x20+10% -negate result.png
            image.Negate();
            image.AdaptiveThreshold(20, 20, new Percentage(10));
            image.Negate();
        }
        #endregion