Gray Level-Median Histogram Based 2D Otsu's Method

Abstract

In this paper, we proposed a gray level-median histogram based two-dimensional (2D) Otsu's method, which is efficient and robust to noise. In our method, we use gray level-median instead of gray level-mean to build 2D histogram. We also proposed a new method to process the edge and noise regions to reduce noise and obtain more clear edges. In order to improve the computation efficiency, our method selects the optimal threshold in each dimension independently using 1D Otsu's method on 1D histogram. Experimental results show that our method has stronger anti-noise capability and faster processing speed compared with the traditional 2D Otsu's method and the traditional 3D Otsu's method. Especially for low SNR image, the segmentation results obtained by our method are much better than those by the traditional 2D Otsu's method and the traditional 3D Otsu's method.