IMAGE ENHANCEMENT USING AUTOMATIC SEGMENTATION TECHNIQUES

This paper provides automatic image segmentation using dynamic region merging algorithm. Image segmentation is to partition the image into meaningful regions with respect to particular application. Image segmentation is basically to partition image into certain number of piec es that have coherent features, and again group the meaningful pieces for the convenience of perceiving. Image segmentation usually serves as the pre - processing before image pattern recognition. Initially the image is clustered and oversegmented using K - me ans and Watershed Algorithm. With an initially oversegmented image in which many regions with homogeneous color are detected, image segmentation is performed by iteratively merging the regions according to novel predicate. Novel predicate is defined by seq uential probability ratio test. Also an faster algorithm is developed to accelerate region merging process which maintains nearest neighbour graph in each iteration of merging to produce segmentation which satisfies global properties.