A survey on neutrosophic medical image segmentation

Abstract

In the last decade, neutrosophic sets (NS), which are defined as the generalization of interval fuzzy sets, have become a hot topic in the computer-vision and machine-learning communities with a number of applications. Researchers in the field of computer vision have applied NS on various image-processing applications. Segmentation is a well-known process in image processing that aims to divide an input image into its regions. While the pixels in a given region should have the same property, the pixels in different regions should have different properties. In this chapter, a survey is presented on NS-based medical image segmentation. As the literature is explored, it is seen that NS-based image segmentation approaches have been applied on various medical images such as breast ultrasounds (BUS), liver computed tomography (CT), brain CTs, dermoscopy, retinal, eye angiography, dental X-rays, etc. Moreover, there have been numerous applications that have used NS in optical image segmentation. Especially, neutrosophic logic has applications in texture image segmentation. In these studies, NS has been generally used for either denoising or image enhancement. Moreover, in most studies, NS has been used for image segmentation. Besides the literature review, several well-known NS-based medical image segmentation approaches are introduced. In these methods, the NS was either used to improve the image quality by contrast enhancement and noise removal or to segment the image into regions of interest and background. The methodologies and results of the investigated methods are given in detail. The general limitations of the NS-based medical image segmentation approaches are also given. The chapter ends with some conclusions and future perspectives. © 2019 Elsevier Inc. All rights reserved.