Abstract
Most traditional superpixel segmentation methods used binary logic to generate superpixels for natural images. When these methods are used for images with significantly fuzzy characteristics, the boundary pixels sometimes cannot be correctly classified. In order to solve this problem, this paper proposes a Superpixel Method Based on Fuzzy Theory (SMBFT), which uses fuzzy theory as a guide and traditional fuzzy c-means clustering algorithm as a baseline. This method can make full use of the advantage of the fuzzy clustering algorithm in dealing with the images with the fuzzy characteristics. Boundary pixels which have higher uncertainties can be correctly classified with maximum probability. The superpixel has homogeneous pixels. Meanwhile, the paper also uses the surrounding neighborhood pixels to constrain the spatial information, which effectively alleviates the negative effects of noise. The paper tests on the images from Berkeley database and brain MR images from the Brain web. In addition, this paper proposes a comprehensive criterion to measure the weights of two kinds of criterions in choosing superpixel methods for color images. An evaluation criterion for medical image data sets employs the internal entropy of superpixels which is inspired by the concept of entropy in the information theory. The experimental results show that this method has superiorities than traditional methods both on natural images and medical images.
Highlights
The definition of superpixel segmentation is that aggregating some pixels together to form the atomic area with certain perceptual significance for replacing the area grid
Owing to the proposed algorithm in this paper based on the fuzzy theory, it can overcome the impact brought by this uncertainty to some extents
Since the data set does not has the ground truth for calculate Undersegmentation error (UE) and Boundary recall (BR), evaluation criterion adopted by the paper does not use the traditional UE and BR
Summary
The definition of superpixel segmentation is that aggregating some pixels together to form the atomic area with certain perceptual significance for replacing the area grid. It can make use of spatial constraint information to be robust to certain noises [1]. The middle level features of the image can be extracted in the atomic area segmented by the superpixel method, which is beneficial for further processing of the image. Some images have the intensity inhomogeneity problem; for example, gray value of white matter in some areas of brain MR image is close to that of gray matter in other areas and even is lower than that of gray matter [3]. As in the atomic area, the contrast ratio of the internal gray values in one superpixel is higher, and there is no intensity inhomogeneity phenomenon in the superpixel, which can avoid the influence of intensity inhomogeneity [4]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
