Learned Image Compression with Separate Hyperprior Decoders

TL;DR

This paper introduces a novel image compression method using three separate hyperprior decoders, leading to more accurate parameter estimation and improved compression performance without increasing computational costs.

Contribution

The paper proposes using three hyperprior decoders to enhance parameter estimation in learned image compression, surpassing previous single-decoder models.

Findings

Achieves 3.36% BD-rate reduction with MS-SSIM optimization.

Maintains negligible impact on coding time and FLOPs.

Outperforms state-of-the-art methods in image compression quality.

Abstract

Learned image compression techniques have achieved considerable development in recent years. In this paper, we find that the performance bottleneck lies in the use of a single hyperprior decoder, in which case the ternary Gaussian model collapses to a binary one. To solve this, we propose to use three hyperprior decoders to separate the decoding process of the mixed parameters in discrete Gaussian mixture likelihoods, achieving more accurate parameters estimation. Experimental results demonstrate the proposed method optimized by MS-SSIM achieves on average 3.36% BD-rate reduction compared with state-of-the-art approach. The contribution of the proposed method to the coding time and FLOPs is negligible.