Multi Resolution Image Segmentation For Scalable Object-Based Wavelet Coding

This paper introduces a multi resolution image segmentation algorithm for scalable object based wavelet coding. This algorithm is based on discrete wavelet transform and multiresolution Markov random field (MMRF) modelling. The major contribution is to match the spatial scalability features of arbitrary shape wavelet transforms with the segmentation algorithm. To optimize the segmentation/extraction of objects/regions of interest in different resolutions of the wavelet pyramid, with scalability constraint, a multi scale analysis is incorporated into the objective function of MMRF segmentation algorithm. The proposed algorithm improves the segmentation result, especially in lower resolutions of the decomposition, over regular multi resolution segmentation in both objective and subjective tests, in yielding an effective segmentation that supports scalable wavelet coding.