Quote: Original post by Timkin
Rofl... everything described here (transforms, filtering, gradient based analysis) is all covered by doing a wavelet analysis of the image. If you're going to do something, don't 'hack it', do it properly. Yes, there are other methods out there that are useful... but they're generally not as robust as this single method.

It would be helpful if you would suggest some books/other resources on the subject. I've found that in subjects that are highly technical (such as wavelets are and wavelet analysis probably is) there is more misinformation on the internet than information, and the information is terse and presumes the reader knows much more math than they do (or at least than I do). Quality books usually don't make as many basic mistakes.

As far as which filter to apply, an anisotophic diffusion of the kind that photoshop provides usually brings out shapes very well. I have no idea how it works, though

Quote: Original post by Steadtler
Timkin, I think a wavelet analysis is overkill for a simple edge detection. Canny have been *proven*, and I mean proven as in mathematically proven, to be more robust and accurate as anything else for step edges detection.

But the problem here is that the OP specifically stated that it wasn't simple edge detection and that he had gradient pencil shading near edges and noise in the image. The pencil shading alone will fool most edge detection algorithms because the shading is made up of many individual lines blended together (but each often has its own tone...depending on the quality of the artist). If they used a crosshatch shading... youch!

Of course there are other algorithms for edge detection... but be wary of claims of robustness... particularly mathematical robustness...because they always rely on assumptions regarding image properties (particularly mathematical proofs).

The proof is in the pudding. If you have the time and knowledge, implement several techniques and choose the one that works the best on a broad range of samples from your problem domain.

As for good books on this topic... just go to your local college library. Most textbooks on Image Processing will have more than enough of what you need. I did have an excellent book on the topic, but it got pinched while at my last job and I cannot for the life of me remember the authors name. 8(