Split-and-merge image segmentation based on localized feature analysis and statistical tests

In this paper, an adaptive split-and-merge image segmentation algorithm based on characteristic features and a hypothesis model is proposed. The analysis of characteristic features provides the requisite parameters that serve as constraints in the hypothesis model. The strength of the proposed method lies in the fact that the parameters in the algorithms are computed automatically and depend only on the context of the image under analysis. One of the key processes, the determination of region homogeneity, is treated as a sequence of decision problems in terms of predicates in the hypothesis model. Experimental results on natural scene pictures and medical images are included to demonstrate the robustness of the algorithm.