JPD-SE: High-Level Semantics for Joint Perception-Distortion Enhancement in Image Compression

TL;DR

This paper introduces a semantic-aware image compression framework that leverages high-level semantics to improve perception quality, rate-distortion performance, and downstream computer vision tasks through a novel joint optimization approach.

Contribution

It proposes a generic framework and training scheme enabling any codec to incorporate high-level semantics for enhanced perception and downstream task performance.

Findings

Semantic-aware codecs outperform traditional codecs in perception quality.

Joint optimization improves rate-distortion-perception trade-offs.

Enhanced downstream computer vision algorithm performance.

Abstract

While humans can effortlessly transform complex visual scenes into simple words and the other way around by leveraging their high-level understanding of the content, conventional or the more recent learned image compression codecs do not seem to utilize the semantic meanings of visual content to their full potential. Moreover, they focus mostly on rate-distortion and tend to underperform in perception quality especially in low bitrate regime, and often disregard the performance of downstream computer vision algorithms, which is a fast-growing consumer group of compressed images in addition to human viewers. In this paper, we (1) present a generic framework that can enable any image codec to leverage high-level semantics and (2) study the joint optimization of perception quality and distortion. Our idea is that given any codec, we utilize high-level semantics to augment the low-level visual features extracted by it and produce essentially a new, semantic-aware codec. We propose a three-phase training scheme that teaches semantic-aware codecs to leverage the power of semantic to jointly optimize rate-perception-distortion (R-PD) performance. As an additional benefit, semantic-aware codecs also boost the performance of downstream computer vision algorithms. To validate our claim, we perform extensive empirical evaluations and provide both quantitative and qualitative results.