A Single Image Compression Framework Combined with Sparse Representation-Based Super-Resolution

DOI

10.2991/esac-15.2015.71How to use a DOI?

Keywords

Image compression, Targeted dictionary training, Super-resolution

Abstract

This paper addresses the problem of single image compression by using sparse representation-based super-resolution modeling. The proposed single image compression framework (SICF) is mainly composed of two components, including image encoder and decoder. Given an image, the framework takes down-sampling processing. Then, the down-sampling image is fed into JPEG and reconstructed by the sparse representation-based super-resolution, which is using the proposed targeted dictionary based on soft-decision adaptive interpolation (TD-SAI) and targeted residual dictionary (TRD). For further improving the quality of the final decoded image, the feedback is designed to obtain residual assistance image which can be compensated for the loss of high-frequency details in process of super-resolution reconstruction. Supported by the targeted dictionary, our method achieved an improvement in objective quality assessment and competitive performance in visual quality when comparing with other SR algorithms. More importantly, the proposed SICF achieved significant bit-rate saving under the same PSNR compared with JPEG standard. Extensive experiments manifest the viability and efficiency of the proposed image compression framework.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Xiaohai He
AU  - Jingbo He
AU  - Jianqiu Huang
AU  - Di Wu
PY  - 2015/08
DA  - 2015/08
TI  - A Single Image Compression Framework Combined with Sparse Representation-Based Super-Resolution
BT  - Proceedings of the 2015 International Conference on Electronic Science and Automation Control
PB  - Atlantis Press
SP  - 292
EP  - 296
SN  - 2352-538X
UR  - https://doi.org/10.2991/esac-15.2015.71
DO  - 10.2991/esac-15.2015.71
ID  - He2015/08
ER  -