Histogram-based Fuzzy C-Means Clustering for Image Binarization

The goal of image binarization is to classify the pixels into black and white correctly. Finding a threshold to binarize the image effectively is the essence in this study. This paper introduces a new algorithm for image binarization based on clustering. The algorithm computes on the histogram and uses the membership partition based on the distance between pixels within local spatial neighbors and clustering centers to accelerate the binarization procedure. Then the weighted factor is introduced to balance the noise-immunity and details. The proposed method combines the global and local ideas in the conventional algorithms. Compared with state-of-the-art algorithms, the proposed algorithm can universally obtain a robust effect for the images within distinct features, especially for the precision images.