Image Thresholding Based on Two-Dimensional Tsallis Gray Entropy Using Fast Iterative Algorithm

Abstract

In order to extract target from complex background more quickly and accurately, and efficiently improve the computational efficiency of the optimal thresholds, the rapid iteration method based on 2D Tsallis gray entropy thresholding is proposed. First of all, the rapid iteration method of one-dimensional gray-entropy threshold Tsallis is proposed. Then considering the within-class grayscale uniformity of image target and background, Tsallis gray entropy thresholding based on Grayscale-Average gray level two-dimensional histogram is derived, which along with the Intermediate variables in the recursion formula can be used to efficiently remove the redundant computation and decrease the romputation. At last, the rapid iteration method of 2D Tsallis gray entropy thresholding is proposed. The efficiency of the romputation is improved greatly by deriving the rorresponding rormula. A great number of experimental results have proven that, compared to the four similar thresholding algorithm, the method put forward in this paper can more precisely and completely fragmenting the target needed with better fragmenting performance and faster running speed, which turns out a real-time and efficient image segmenting method.