Minimum Cumulative Residual Information Energy Thresholding on Circular Histogram
Jing Liu, Jiulun Fan, Jinjing Ai · 2023
Shannon used logarithmic operation and probability distribution function to define a way to measure information uncertainty: Shannon entropy. Rao used cumulative distribution function to replace the probability distribution function of Shannon entropy to define another way to measure information uncertainty: cumulative residual entropy. Corresponding to Shannon entropy, Onicescu defined a way to measure the certainty of information: information energy. In view of these, cumulative residual information energy, a new way to measure the certainty of information is proposed and used to thresholding segmentation based on H component in the HSI color model. Since the H component of HSI color model is a circular histogram, a recursive algorithm based on the periodic characteristics of circular histogram is given, which improves the computational efficiency of the proposed thresholding segmentation method. Compared with the related thresholding methods on the circular histogram, the experimental results show that the proposed method is effective.