Rate-Distortion Optimized Sparse Coding with Ordered Dictionary for Image Set Compression

Xinfeng Zhang,Weisi Lin,Yabin Zhang,Shiqi Wang,Siwei Ma,Lingyu Duan,Wen Gao

DOI: https://doi.org/10.1109/tcsvt.2017.2748382

IF: 5.859

2017-01-01

IEEE Transactions on Circuits and Systems for Video Technology

Abstract:Image set compression has recently emerged as an active research topic due to the rapidly increasing demand in cloud storage. In this paper, we propose a novel framework for image set compression based on the rate-distortion optimized sparse coding. Specifically, given a set of similar images, one representative image is first identified according to the similarity among these images, and a dictionary can be learned subsequently in wavelet domain from the training samples collected from the representative image. In order to improve coding efficiency, the dictionary atoms are reordered according to their use frequencies when representing the representative image. As such, the remaining images can be efficiently compressed with sparse coding based on the reordered dictionary that is highly adaptive to the content of the image set. To further improve the efficiency of sparse coding, the number of dictionary atoms for image patches is further optimized in a rate-distortion sense. Experimental results show that the proposed method can significantly improve the image compression performance compared with JPEG, JPEG2000, and the state-of-the-art dictionary learning-based methods.