ON THE APPLICATION OF GIBBS RANDOM-FIELD IN IMAGE-PROCESSING - FROM SEGMENTATION TO ENHANCEMENT

The Gibbs random field (GRF) has proved to be a simple and practical w ay of parameterizing the Markov random field which has been widely use d to model an image or image-related process in many image processing applications, In particular, the GRF can be employed to construct an e fficient Bayesian estimation that often yields optimal results, We des cribe how the GRF can be efficiently incorporated into optimization pr ocesses in several representative applications, ranging from image seg mentation to image enhancement One example is the segmentation of comp uterized tomography (CT) volumetric image sequence in which the GRF ha s been incorporated into K-means clustering to enforce the neighborhoo d constraints. Another example is the artifact removal in discrete cos ine transform-based low bit rate image compression where GRF has been used to design an enhancement algorithm that reduces the ''blocking ef fect'' and the ''ringing effect'' while still preserving the image det ails, The third example is the integration of GRF in a wavelet-based s ubband video coding scheme in which the high-frequency subbands are se gmented and quantized with spatial constraints specified by a GRF, and the subsequent enhancement of the decompressed images is accomplished by smoothing with another type of GRF. With these diverse examples, w e are able to demonstrate that various features of images can all be p roperly characterized by a GRF. The specific form of the GRF can be se lected according to the characteristics of an individual application. We believe that the GRF is a powerful tool to exploit the spatial depe ndency in various images, and is applicable to many image processing t asks.