I am looking for good methods, preferably parameterless, to detect textured areas in an image. I mean random textures, not regular patterns. The final goal is to discriminate between significant edges and false edges in texture, which can have a significant intensity.

In my opinion, a discriminative characteristic of textured areas is that the gradient direction can strongly from pixel to pixel, but I have found no good way to rate this. In particular, the random behavior makes local estimators fairly unstable.

I am aware of the technique of surround suppression proposed by Grigorescu, Petkov et al., but I would prefer a detection technique that works independently of the presence of true edges, such as a texturedness indicator. I also know the interesting paper "Detection of Textured Areas in Images Using a Disorganization Indicator Based on Component Counts, Ruth Bergman, H. Nachlieli, G. Ruckenstein", which goes in a nice direction.

Are you aware of alternative techniques ?

Please note that my question is not about texture edges (boundary between different textures).


Your Answer

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

Browse other questions tagged or ask your own question.