Spinal Cord Image Processing Using Enhanced Multi-Scale Retinex Algorithm

Abstract

This paper proposes a development of novel enhanced multiscale retinex algorithm for medical image enhancement in light of multi-rate testing. The proposed work embraces HSV color space, since HSV color space separates color from power. The enhancement of medical image is accomplished by down testing the original image into five forms, in particular, minuscule, little, medium, fine, and typical scale. Further, the contrast stretching and multiscale retinex (MSR) techniques are taken advantage of to improve the scaled forms of the image. At last, the image is reconstructed by joining every one of these scales in a efficient manner to get the composite enhanced image. The efficiency of the proposed algorithm is approved by directing various experiments on spinal cord images. Reconstructed image utilizing proposed technique features the subtleties (edges and tissues), decreases image noise (Gaussian and Speckle) and works on the general difference. The proposed algorithm likewise improves sharp edges of the tissue encompassing the spinal cord regions which is valuable for diagnosis of spinal cord lesions.