A novel blind source separation technique using fractional Fourier transform for denoising medical images

In this paper we propose a novel algorithm to denoise medical images. The proposed method provides enhanced visual clarity to experts for utilizing such images during diagnosis. Medical images are unique in the fact that they are often low contrast and extremely noisy in nature because of the circumstances under which they are captured. Denoising these images without a priori knowledge of the noise PDF is always difficult and successful denoising of such images should not induce other artifacts or blurring of edges in the images. The proposed algorithm utilizes the technique of blind source separation (BSS) and the fractional Fourier transform to provide superior and stable denoising of medical images. The performance of the algorithm is compared with the recent genetic algorithm based wavelet thresholding method and the results are presented.