Expandable, Compressible, Mineable: Open-World Thermal Image Restoration

2026-05-16

AI Summary

Existing thermal infrared image restoration methods are constrained by the closed-world assumption, rendering them ineffective against novel degradation patterns continually emerging in open environments. To address this limitation, this work proposes ECMRNet, a framework that enables continual learning through an “expand-compress-mine” closed-loop mechanism. It employs a group-isolated subspace architecture to support parameter isolation and rapid model expansion, incorporates structural entropy-based pruning for adaptive compression, and introduces a sub-degradation knowledge mining module to enhance restoration performance under complex, composite degradations. Extensive experiments demonstrate that ECMRNet significantly outperforms current state-of-the-art methods across diverse single and composite degradation scenarios while substantially reducing both model parameters and computational overhead.

Abstract

In open-world settings, thermal infrared (TIR) image degradations continuously emerge and evolve, while most existing all-in-one restoration methods are built on a closed-set assumption and struggle to continually adapt to novel degradations. To address this, we propose ECMRNet, an Expandable, Compressible, and Mineable Restoration Network for open-world TIR restoration from a continual learning perspective. Conceptually, ECMRNet unifies continual degradation learning as an "expand-compress-mine" closed-loop process, enabling sustained adaptation to new degradations with controllable evolution. Structurally, ECMRNet decomposes intermediate representations into group-isolated subspaces, and achieves strict parameter isolation and fast adaptation to new degradations by freezing historical groups and isomorphically expanding new ones. To curb model growth as tasks accumulate, we present Structural Entropy Pruning, which identifies and removes redundant channel groups via two-dimensional structural entropy minimization, achieving information contribution-driven adaptive compression. Moreover, we design a Sub-degradation Knowledge Mining Module that dynamically retrieves and recombines transferable components from historical representations to improve restoration under compound degradations. Experimental results demonstrate that ECMRNet achieves superior overall performance across diverse single and compound degradations while using fewer parameters and lower computational cost. The source code is available at https://github.com/Kust-lp/ECMRNet.

Problem

• thermal infrared image restoration
• open-world
• continual learning
• degradation adaptation
• model expandability

Innovation

• continual learning
• expandable architecture
• structural entropy pruning
• parameter isolation
• knowledge mining