Color Image Enhancement Using Multiscale Retinex with Guided Filter

Abstract

Image enhancement algorithms are developed to achieve improved image quality. This paper presents a novel technique to enhance colored images using Multi Scale Retinex (MSR) along with the Guided Filter (GF). The proposed technique has several steps. Initially, the PCA isolates the input RGB image into the luminance and chrominance channels. On luminance channel, the MSR is applied in three steps. Meanwhile, the GF is introduced, which divides the image into high and low frequency components. Finally, Contrast Stretching (CS) is applied that expands the contrast of the image. Proposed method yields visually pleasing images on NASA retinex dataset and outperforms few techniques in terms of the entropy, the PSNR, and memory usage. In addition, the proposed method is computationally efficient.