Effect of using Genetic Algorithm to denoise MRI images corrupted with Rician Noise

It is well known that Genetic Algorithm (GA) uses large number of solutions, instead of a single solution for searching. This brings an important part to the robustness of genetic algorithms. It improves the chance of reaching the global optimum and nearly unbiased optimization techniques for sampling a large solution space. GA adapted in image processing because of this unbiased stochastic sampling. In this paper GA is proposed for removal of Rician Noise. This kind of noise mainly occurs in low signal to noise (SNR) regions. True low signal is corrupted due to presence of Rician noise and measurement gets hampered in low SNR regions. Noise in magnetic resonance (MR) magnitude image maintains Rician distribution. It is a signal dependent Noise. To overcome this problem real and imaginary data in the image field are rectified, before construction of the magnitude image. The noise-reduction filtering (or denoising) is accomplished by Genetic Algorithm. A fresh genetic operator is used that combines crossover and adaptive mutation to improve the convergence rate and solution quality of GA. The proposed technique effectively reduces the standard deviation and significantly lowers the rectified noise.