Color-Embedded-Grayscale Quality Enhancement Using Deep Convolutional Networks

The color recovery process of color-embedded-grayscale image can be considered as ill-posed inverse imaging problem. In most cases, the quality of recovered color image is less satisfied due to color distortion and checkerboard artifacts. To overcome this aforementioned problem, we propose a deep learning-based method for increasing the quality of recovered color image. This method employs the Convolutional Neural Networks (CNN) as post processing step in the color recovery stage. As reported in the Experimental Section, the proposed method yields the promising results for the color recovery task examined over subjective and objective quality assessments.