DeLightMono: Enhancing Self-Supervised Monocular Depth Estimation in Endoscopy by Decoupling Uneven Illumination

2025-11-25

AI Summary

Monocular depth estimation in endoscopic imaging suffers significant performance degradation under non-uniform illumination—particularly in low-light regions. Method: This paper proposes the first illumination-decoupled self-supervised depth estimation framework tailored for endoscopic scenes. Grounded in an illumination-reflectance-depth decomposition model, it introduces an auxiliary decomposition network and a joint optimization mechanism, along with a novel self-supervised loss function explicitly guided by illumination–reflectance decoupling—requiring no additional annotations or illumination-correction data. Contribution/Results: It is the first work to explicitly model illumination components within monocular depth learning, effectively suppressing low-light interference. Experiments on two public endoscopic datasets demonstrate substantial improvements in depth accuracy and robustness, achieving superior overall metrics compared to state-of-the-art methods. Ablation studies confirm the effectiveness of each component.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Self-supervised monocular depth estimation serves as a key task in the development of endoscopic navigation systems. However, performance degradation persists due to uneven illumination inherent in endoscopic images, particularly in low-intensity regions. Existing low-light enhancement techniques fail to effectively guide the depth network. Furthermore, solutions from other fields, like autonomous driving, require well-lit images, making them unsuitable and increasing data collection burdens. To this end, we present DeLight-Mono - a novel self-supervised monocular depth estimation framework with illumination decoupling. Specifically, endoscopic images are represented by a designed illumination-reflectance-depth model, and are decomposed with auxiliary networks. Moreover, a self-supervised joint-optimizing framework with novel losses leveraging the decoupled components is proposed to mitigate the effects of uneven illumination on depth estimation. The effectiveness of the proposed methods was rigorously verified through extensive comparisons and an ablation study performed on two public datasets.

Problem

• Addressing uneven illumination in endoscopic depth estimation
• Decoupling illumination effects to improve depth accuracy
• Developing self-supervised framework for low-light medical imaging

Innovation

• Decouples illumination from reflectance and depth
• Uses self-supervised joint-optimizing framework with novel losses
• Employs auxiliary networks for image decomposition