RIDE: Retinex-Informed Decoupling for Exposing Concealed Objects
The paper introduces RIDE, a novel approach for Concealed Object Segmentation (COS) using Retinex theory. This method focuses on homogeneous image decomposition to enhance the visibility of concealed objects by separating illumination and reflectance components. The authors propose a framework that includes a decomposition module, an attention mechanism, and a contrastive loss to improve segmentation performance across various tasks.
- ▪RIDE utilizes Retinex theory for homogeneous image decomposition in Concealed Object Segmentation.
- ▪The method separates illumination and reflectance components to improve object visibility.
- ▪It includes a Task-Driven Retinex Decomposition module and a Discriminability Gap Attention mechanism.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Computer Vision and Pattern Recognition arXiv:2605.15450 (cs) [Submitted on 14 May 2026] Title:RIDE: Retinex-Informed Decoupling for Exposing Concealed Objects Authors:Chunming He, Rihan Zhang, Dingming Zhang, Chengyu Fang, Longxiang Tang, Jingjia Feng, Fengyang Xiao, Sina Farsiu View a PDF of the paper titled RIDE: Retinex-Informed Decoupling for Exposing Concealed Objects, by Chunming He and 7 other authors View PDF HTML (experimental) Abstract:Concealed Object Segmentation (COS) encompasses a family of dense-prediction tasks, including camouflaged object detection, polyp segmentation, transparent object detection, and industrial defect inspection, where targets are visually entangled with their surroundings through different physical mechanisms.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
