Dynamic Stochastic Resonance Based Diffusion-Weighted Magnetic Resonance Image Enhancement Using Multi-Objective Particle Swarm Optimization

• Published in: Journal of medical and biological engineering Vol. 36; no. 6; pp. 891 - 900
• Main Authors: Singh, Munendra, Sharma, Neeraj, Verma, Ashish, Sharma, Shiru
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2016; Springer Nature B.V
• Subjects: Biomedical Engineering and Bioengineering; Cell Biology; Diffusion; Engineering; Image contrast; Image enhancement; Image quality; Imaging; Magnetic resonance imaging; Medical imaging; Multiple objective analysis; Neuroimaging; Optimization algorithms; Original Article; Parameters; Particle swarm optimization; Radiology; Signal to noise ratio; Stochastic resonance; Multi-objective particle swarm optimization; Diffusion-weighted MR image; Dynamic stochastic resonance; Optimum contrast; Mean opinion score; Contrast enhancement factor
• ISSN: 1609-0985; 2199-4757
• DOI: 10.1007/s40846-016-0186-0

Diffusion weighted (DW) magnetic resonance (MR) imaging maps the diffusion process of water in the tissues. DW-MR image is useful to probe the tissue microstructure, but suffers from inherent low signal to noise ratio and poor contrast. Dynamic stochastic resonance (DSR) utilizes the noise to enhance the low contrast image where the quality of the processed image depends on the bistability parameters of DSR and the number of iterations. This paper presents an approach that optimally finds the bistability parameters and number of iterations for the maximization of competitive image quality indices: contrast enhancement factor and mean opinion score using multi-objective particle swarm optimization. The proposed Particle Swarm Optimization optimized DSR algorithm has been tested on 40 DW-MR brain images of different subjects. The quantified results show average contrast enhancement factor, 1.603 and average perceptual quality measure, 9.508. These values are significantly higher than image quality indices of original image, the images that are produced by conventional enhancement methods and filtering followed by enhancement methods.