Two-dimensional Kapur Entropy Multilevel Thresholding Algorithm Based on Dynamic Programming and Spatial Contextual Information

The Kapur entropy multilevel thresholding image segmentation algorithm based on the two-dimensional histogram does not consider the spatial contextual information of image pixels, and the algorithm's time complexity increases exponentially with the increase of the number of thresholds. To address the above problems, this paper proposes a 2D-Kapur entropy multilevel thresholding image segmentation algorithm based on an energy curve by using both the neighborhood information and spatial contextual information of pixels. To improve the computational efficiency of the multilevel thresholding segmentation algorithm, a dynamic programming computational framework of the 2D-Kapur entropy multilevel thresholding algorithm based on the energy curve is constructed. The performance of the proposed algorithm is compared with several different image thresholding algorithms on the noise and complex images using several evaluation indexes. The experimental results verify the effectiveness of the proposed algorithm.