Efficient Degradation-aware Any Image Restoration

TL;DR

DaAIR is an efficient, unified image restoration model that dynamically adapts to various degradations using a degradation-aware embedding, outperforming existing methods while maintaining computational efficiency.

Contribution

We introduce DaAIR, a novel degradation-aware image restorer that dynamically allocates model capacity and employs a cost-efficient update mechanism for improved performance.

Findings

DaAIR outperforms state-of-the-art All-in-One models.

DaAIR surpasses degradation-specific methods in diverse restoration tasks.

The model maintains high efficiency with reduced computational overhead.

Abstract

Reconstructing missing details from degraded low-quality inputs poses a significant challenge. Recent progress in image restoration has demonstrated the efficacy of learning large models capable of addressing various degradations simultaneously. Nonetheless, these approaches introduce considerable computational overhead and complex learning paradigms, limiting their practical utility. In response, we propose \textit{DaAIR}, an efficient All-in-One image restorer employing a Degradation-aware Learner (DaLe) in the low-rank regime to collaboratively mine shared aspects and subtle nuances across diverse degradations, generating a degradation-aware embedding. By dynamically allocating model capacity to input degradations, we realize an efficient restorer integrating holistic and specific learning within a unified model. Furthermore, DaAIR introduces a cost-efficient parameter update mechanism that enhances degradation awareness while maintaining computational efficiency. Extensive comparisons across five image degradations demonstrate that our DaAIR outperforms both state-of-the-art All-in-One models and degradation-specific counterparts, affirming our efficacy and practicality. The source will be publicly made available at https://eduardzamfir.github.io/daair/