Exploring Spectral Characteristics for Single Image Reflection Removal

2025-09-15

AI Summary

Image reflection removal is an ill-posed inverse problem, primarily due to the strong coupling between reflection and transmission layers in a single image and the lack of effective modeling of their distinct spectral responses. To address this, we propose a novel spectrum-aware reflection removal paradigm. First, we construct a spectral codebook to explicitly model the wavelength-selective nature of reflected light. Second, we design a dual-spectral prior module to enforce physical consistency and introduce a spectrum-aware Transformer that jointly models spatial and spectral (wavelength) dimensions for effective layer decoupling. Our method integrates spectral reconstruction, adaptive spectral enhancement, and cross-domain feature co-optimization. Extensive experiments on three benchmark datasets demonstrate that our approach significantly outperforms existing state-of-the-art methods, achieving superior reflection removal accuracy and enhanced generalization capability.

Technology Category

• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Responsible Web: Algorithmic accountability and transparency on the web
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Eliminating reflections caused by incident light interacting with reflective medium remains an ill-posed problem in the image restoration area. The primary challenge arises from the overlapping of reflection and transmission components in the captured images, which complicates the task of accurately distinguishing and recovering the clean background. Existing approaches typically address reflection removal solely in the image domain, ignoring the spectral property variations of reflected light, which hinders their ability to effectively discern reflections. In this paper, we start with a new perspective on spectral learning, and propose the Spectral Codebook to reconstruct the optical spectrum of the reflection image. The reflections can be effectively distinguished by perceiving the wavelength differences between different light sources in the spectrum. To leverage the reconstructed spectrum, we design two spectral prior refinement modules to re-distribute pixels in the spatial dimension and adaptively enhance the spectral differences along the wavelength dimension. Furthermore, we present the Spectrum-Aware Transformer to jointly recover the transmitted content in spectral and pixel domains. Experimental results on three different reflection benchmarks demonstrate the superiority and generalization ability of our method compared to state-of-the-art models.

Problem

• Removing reflections from single images
• Separating overlapping reflection and transmission components
• Addressing spectral property variations in reflection removal

Innovation

• Spectral Codebook reconstructs optical spectrum
• Spectral prior modules redistribute spatial pixels
• Spectrum-Aware Transformer jointly recovers content