16 - Optimization-based neutrosophic set in computer-aided diagnosis

Abstract

Computer-aided detection (CADe) and diagnosis (CADx) have a great impact in medical images interpretation and diagnostic decisions in clinical practical situations. However, the development of CAD systems generated numerous challenging complications for both the mathematical community and the medical computing domain. Uncertainty, vagueness, and imprecision are critical limitations in the medical information leading to poor diagnosis quality. Recently, neutrosophic theory played an interesting role in supporting CAD systems by handling the limitations of the medical images characteristics. The previous studies proved the significant role of neutrosophic sets (NS), as a powerful neutrosophic theory model, in the different medical image processing stages, including denoising, enhancement, and segmentation. Additionally, on the other side, several optimization algorithms have proved their efficiency in medical applications. Accordingly, this chapter introduces the framework for integrating the NS with the optimization algorithms. Firstly, the main concepts of the NS are introduced to explain the generation of the neutrosophic image. Moreover, highlights of optimization techniques are provided, followed by the proposal of a general framework for integrating NS and any optimization technique, called OptNS in the present chapter. Furthermore, an extensive discussion on the different integrated optimization and NS models in CAD systems is included. The reported studies proved the impact of integrating the neutrosophic model with optimization algorithms on the diagnosis accuracy which adopted auspicious aspirants for more advanced CAD systems. For future research studies on OptNS for CAD, it is recommended to work on the optimization of the NS operators as well as optimizing the different parameters in the different medical image processing to gain the NS benefits of both the NS and optimization.