Colorization Based Compression Using Mean Shift Segmentation and Sparse Recovery Algorithm

In this paper, the colorization-based coding problem has been solved using smooth L0 (SL0) and L1 norm (OMP) minimization sparse recovery algorithms. In colorization-based coding method at the encoder, only few representative pixels (RP) for the chrominance values are sent along with luminance part of image to the decoder where the chrominance values for all the pixels are reconstructed by colorization method. The major issue in colorization-based coding method is to select or extract the RP well so as to reconstruct chrominance information of all the pixels, which is required for the reconstruction of good quality color image. Colorization matrix is constructed using mean-shift segmentation of luminance part of the image which is used in extracting RP set by applying OMP and SL0 algorithm. The size of transmitted image is reduced because of reduced number of RP set. The quality of reconstructed image has been evaluated using MSE, PSNR and SSIM parameters.