Development of multiscale retinex algorithm for medical image enhancement based on multi-rate sampling

This paper proposes a development of novel multiscale retinex algorithm for medical image enhancement based on multi-rate sampling. The proposed work adopts HSV color space, since HSV color space separates color from intensity. The enhancement of medical image is achieved by downsampling the original image into five versions, namely, tiny, small, medium, fine, and normal scale. Further, the contrast stretching and multiscale retinex (MSR) techniques are exploited in order to enhance the scaled versions of the image. Finally, the image is reconstructed by combining each of these scales in an efficient way to obtain the composite enhanced image. The efficiency of the proposed algorithm is validated by conducting numerous experiments on spinal cord images. Reconstructed image using proposed method highlights the details (edges and tissues), reduces image noise (Gaussian and Speckle) and improves the overall contrast. The proposed algorithm also enhances sharp edges of the tissue surrounding the spinal cord regions which is useful for diagnosis of spinal cord lesions. Elaborated experiments are conducted on several medical images and results presented show that the enhanced medical pictures are of good quality and is far better compared to other researcher methods.