Research

Paper

TESTING March 06, 2026

SLER-IR: Spherical Layer-wise Expert Routing for All-in-One Image Restoration

Authors

Peng Shurui, Xin Lin, Shi Luo, Jincen Ou, Dizhe Zhang, Lu Qi, Truong Nguyen, Chao Ren

Abstract

Image restoration under diverse degradations remains challenging for unified all-in-one frameworks due to feature interference and insufficient expert specialization. We propose SLER-IR, a spherical layer-wise expert routing framework that dynamically activates specialized experts across network layers. To ensure reliable routing, we introduce a Spherical Uniform Degradation Embedding with contrastive learning, which maps degradation representations onto a hypersphere to eliminate geometry bias in linear embedding spaces. In addition, a Global-Local Granularity Fusion (GLGF) module integrates global semantics and local degradation cues to address spatially non-uniform degradations and the train-test granularity gap. Experiments on three-task and five-task benchmarks demonstrate that SLER-IR achieves consistent improvements over state-of-the-art methods in both PSNR and SSIM. Code and models will be publicly released.

Metadata

arXiv ID: 2603.05940
Provider: ARXIV
Primary Category: cs.CV
Published: 2026-03-06
Fetched: 2026-03-09 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.05940v1</id>\n    <title>SLER-IR: Spherical Layer-wise Expert Routing for All-in-One Image Restoration</title>\n    <updated>2026-03-06T06:16:57Z</updated>\n    <link href='https://arxiv.org/abs/2603.05940v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.05940v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Image restoration under diverse degradations remains challenging for unified all-in-one frameworks due to feature interference and insufficient expert specialization. We propose SLER-IR, a spherical layer-wise expert routing framework that dynamically activates specialized experts across network layers. To ensure reliable routing, we introduce a Spherical Uniform Degradation Embedding with contrastive learning, which maps degradation representations onto a hypersphere to eliminate geometry bias in linear embedding spaces. In addition, a Global-Local Granularity Fusion (GLGF) module integrates global semantics and local degradation cues to address spatially non-uniform degradations and the train-test granularity gap. Experiments on three-task and five-task benchmarks demonstrate that SLER-IR achieves consistent improvements over state-of-the-art methods in both PSNR and SSIM. Code and models will be publicly released.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <published>2026-03-06T06:16:57Z</published>\n    <arxiv:primary_category term='cs.CV'/>\n    <author>\n      <name>Peng Shurui</name>\n    </author>\n    <author>\n      <name>Xin Lin</name>\n    </author>\n    <author>\n      <name>Shi Luo</name>\n    </author>\n    <author>\n      <name>Jincen Ou</name>\n    </author>\n    <author>\n      <name>Dizhe Zhang</name>\n    </author>\n    <author>\n      <name>Lu Qi</name>\n    </author>\n    <author>\n      <name>Truong Nguyen</name>\n    </author>\n    <author>\n      <name>Chao Ren</name>\n    </author>\n  </entry>"
}