This paper proposes an automatic global thresholding method based on 2D Tsallis-Havrda-Charva+-t entropy and histogram of Local Binary Patterns (LBP). Tsalli-Havrda-Charvat entropy is obtained from 2D histogram, which has determined by using the LBP decimal value and the average decimal value of its neighborhood. Based on this entropy we obtain the optimal threshold pair by maximizing the criterion function. LBP histogram is adopted to capture the texture information. LBP's high performance for texture characterization helps to make our method more suitable for thresholding the images in problem. In this paper we report the effectiveness of our thresholding method when applied to some real-world and synthetic images, and experiments show that the performance of our proposed method is promising, robust and fast.
Keywords
Image Segmentation, 2D Histogram, Local Binary Pattern, Thresholding.
User
Font Size
Information