Spatial Competition for Low-Complexity Learned Image Compression

TL;DR

This paper presents a low-complexity learned image compression method that uses spatial competition among neural codecs, achieving high performance with reduced decoding complexity and fast encoding.

Contribution

It introduces a novel spatial competition framework with per-region codec selection guided by a mode map, enabling efficient and adaptive image compression.

Findings

Achieves up to -14.5% rate reduction over a single codec.

Reaches HEVC-level performance with 1433 MACs per pixel decoding complexity.

Enables fast encoding and low-complexity decoding.

Abstract

Autoencoder-based image codecs achieve state-of-the-art compression performance but often incur high computational complexity, particularly at decoding time. This work introduces a low-complexity learned image compression framework based on spatial competition between multiple specialized neural codecs. For each image region, the encoder selects the codec that best matches the local content according to a rate-distortion cost. A mode map is transmitted as side information to indicate the per-region codec selection. At decoding time, this mode map-based selection guides reconstruction while preserving the complexity of a single codec. This design enables per-image adaptation with low decoding complexity and fast encoding. On the CLIC 2020 dataset, our method achieves up to -14.5% rate reduction compared to a single codec and reaches HEVC-level performance with a decoding complexity of 1433 MACs per pixel.