Image filtering and segmentation using robust structure-adaptive estimation of intensity

A structure-adaptive approach to the evaluation of image intensity for adaptive filtering or segmentation is described in the context of a two-region structural image model. The proposed adaptive estimation procedure is based on the selection of the best fitting structuring region from all available multiple structuring regions by the maximum a posteriori probability principle. In application to image filtering, the described method allows one to suppress noise and, at the same time, not damage the initial image including corner edges and some image fine details. It provides robust binary segmentation of local objects of interest or their edges on a noisy background. The proposed approach has been tested on different real images including radiographic images from medical and industrial diagnostics