Contrast-Prior Enhanced Duality for Mask-Free Shadow Removal

2025-07-29

AI Summary

Existing shadow removal methods heavily rely on manually annotated or estimated shadow masks, limiting their practical applicability. To address the ambiguity between local contrast cues and low-albedo objects/textures in complex scenes, we propose a mask-free shadow removal framework featuring an Adaptive Gated Dual-branch Attention (AGBA) mechanism that dynamically recalibrates contrast signals. Furthermore, we introduce a Diffusion-based Frequency-domain–Contrast Fusion Network (FCFN), which jointly models global structural information in the frequency domain and local spatial details, enabling soft boundary preservation and faithful texture recovery. To our knowledge, this is the first work to holistically integrate contrast priors, dual-branch attention, frequency-domain feature fusion, and diffusion-inspired generation within a mask-free paradigm. Our method achieves state-of-the-art performance without shadow masks—matching the quality of mask-guided approaches—while significantly improving robustness and visual fidelity in challenging shadow scenarios.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Adversarial Learning & Robustness
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Existing shadow removal methods often rely on shadow masks, which are challenging to acquire in real-world scenarios. Exploring intrinsic image cues, such as local contrast information, presents a potential alternative for guiding shadow removal in the absence of explicit masks. However, the cue's inherent ambiguity becomes a critical limitation in complex scenes, where it can fail to distinguish true shadows from low-reflectance objects and intricate background textures. To address this motivation, we propose the Adaptive Gated Dual-Branch Attention (AGBA) mechanism. AGBA dynamically filters and re-weighs the contrast prior to effectively disentangle shadow features from confounding visual elements. Furthermore, to tackle the persistent challenge of restoring soft shadow boundaries and fine-grained details, we introduce a diffusion-based Frequency-Contrast Fusion Network (FCFN) that leverages high-frequency and contrast cues to guide the generative process. Extensive experiments demonstrate that our method achieves state-of-the-art results among mask-free approaches while maintaining competitive performance relative to mask-based methods.

Problem

• Removing shadows without relying on shadow masks
• Distinguishing shadows from low-reflectance objects and textures
• Restoring soft shadow boundaries and fine-grained details

Innovation

• Adaptive Gated Dual-Branch Attention mechanism
• Diffusion-based Frequency-Contrast Fusion Network
• Dynamic filtering and re-weighting contrast prior