2
$\begingroup$

I have been scratching my head for a while. What I have is a scanned PDF document with text and water marked logo at the back as in the below image.

enter image description here

I want to do OCR over this, which becomes very difficult because of the logo. All the ratchet I've done so far is for coloured images where they can find contrast difference. I've hit a wall when solving the same for an B&W image as shown. Would love any help/suggestions for an approach/method to achieve what I am looking for.

$\endgroup$
0
$\begingroup$

Take a look at the python cv2 module. It has functions that should enable you to remove the watermark. If you have a separate image of the watermark and it is always in the same coordinate location in each image you should be able to subtract it from the images

$\endgroup$
1
  • $\begingroup$ Would that leave the text legible? I mean, if all the watermark is gone, but the watermark itself is on top of some text and you remove all that from the image, you'd be losing information of interest. $\endgroup$ – 89f3a1c Jan 31 at 4:00
0
$\begingroup$

Before we delve deeper, is the watermark like a light gray watermark or is it dark like the text? Because it could be as simple as thresholding on some grays cake value to filter it out.

If it is dark like the text, I'm not 100% sure if this would work, but you could build a small application to create a dataset of such images where you superimpose some watermark over the text. After which you could train a style-gan like model on the dataset to transform the image and clean up the text. Anybody more experienced with GANs should be able to confirm if this approach is worthwhile.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.