How can I segment text from images with complex backgrounds?
I've tried k-means, but it picks up a lot of noise. Any ideas on what else I can try? Here are examples of images https://imgur.com/a/L7eyMUJ
please show your code and put some example image here, not on an external bin, where it will expire, and render your question somewhat useless, thank you.
the images you posted are CAPTCHAs. you won't get free solutions here. I will tell you that you'll need deep learning in general.