EVC: Towards Real-Time Neural Image Compression with Mask Decay

TL;DR

This paper introduces EVC, a real-time neural image compression model that outperforms traditional codecs like VVC in rate-distortion performance while maintaining high speed, and features scalable encoding with dynamic complexity.

Contribution

The paper presents a single-model variable-bit-rate neural image codec with mask decay and residual learning, enabling real-time performance and scalable complexity.

Findings

EVC achieves 30 FPS on 768x512 images, outperforming VVC in RD performance.

Small models reach 30 FPS on 1920x1080 images with competitive RD performance.

Mask decay and residual learning significantly improve the scalability and performance of neural image compression.

Abstract

Neural image compression has surpassed state-of-the-art traditional codecs (H.266/VVC) for rate-distortion (RD) performance, but suffers from large complexity and separate models for different rate-distortion trade-offs. In this paper, we propose an Efficient single-model Variable-bit-rate Codec (EVC), which is able to run at 30 FPS with 768x512 input images and still outperforms VVC for the RD performance. By further reducing both encoder and decoder complexities, our small model even achieves 30 FPS with 1920x1080 input images. To bridge the performance gap between our different capacities models, we meticulously design the mask decay, which transforms the large model's parameters into the small model automatically. And a novel sparsity regularization loss is proposed to mitigate shortcomings of $L_p$ regularization. Our algorithm significantly narrows the performance gap by 50% and 30% for our medium and small models, respectively. At last, we advocate the scalable encoder for neural image compression. The encoding complexity is dynamic to meet different latency requirements. We propose decaying the large encoder multiple times to reduce the residual representation progressively. Both mask decay and residual representation learning greatly improve the RD performance of our scalable encoder. Our code is at https://github.com/microsoft/DCVC.