Improved quality of JPEG compressed image using approximate K-SVD algorithm

JPEG compression scheme is one of the most popular compression techniques which are used frequently in the real world for compression. This is used to reduce the memory consumption while forwarding the image from one place to another place. The quality of images is reduced compared to original image after decompression especially if the compression ratio is high. There are many methods that reduce the artifacts produced by JPEG compression. The K-SVD method developed to reduce the artifacts present in the image after decompression is quite computationally demanding, especially when the dimensions of the dictionary increase or the number of training signals becomes large. Proposed work introduces a new methodology called Approximate K-SVD which will improve the explicit SVD computation. The two basic components of the proposed work are the replacement of the exact SVD computation with a much quicker approximation. An efficient implementation of the algorithm which reduces its complexity as well as its memory requirements. This work improves the quality of images in terms of PSNR and SSIM along with low computational complexity.