Saliency Driven Perceptual Image Compression
Preprint
Abstract
This paper proposes a new end-to-end trainable model for lossy image compression, which includes several novel components. The method incorporates 1) an adequate perceptual similarity metric; 2) saliency in the images; 3) a hierarchical auto-regressive model. This paper demonstrates that the popularly used evaluations metrics such as MS-SSIM and PSNR are inadequate for judging the performance of image compression techniques as they do not align with the human perception of similarity. Alternatively, a new metric is proposed, which is learned on perceptual similarity data specific to image compression. The proposed compression model incorporates the salient regions and optimizes on the proposed perceptual similarity metric. The model not only generates images which are visually better but also gives superior performance for subsequent computer vision tasks such as object detection and segmentation when compared to existing engineered or learned compression techniques.Keywords
All Related Versions
This publication has 11 references indexed in Scilit:
- Generative Adversarial Networks for Extreme Learned Image Compression2019
- The 2018 PIRM Challenge on Perceptual Image Super-Resolution2019
- Conditional Probability Models for Deep Image Compression2018
- The Perception-Distortion Tradeoff2018
- Deeply Supervised Salient Object Detection with Short ConnectionsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2018
- Cut, Paste and Learn: Surprisingly Easy Synthesis for Instance Detection2017
- Deep Residual Learning for Image Recognition2016
- Microsoft COCO: Common Objects in Context2014
- Context-based adaptive binary arithmetic coding in the H.264/AVC video compression standardIEEE Transactions on Circuits and Systems for Video Technology, 2003
- The JPEG 2000 still image compression standardIEEE Signal Processing Magazine, 2001