Single Image Reflection Removal via inter-layer Complementarity
This work addresses the challenge of removing reflections from single images, which is important for applications like photography and computer vision, but it is incremental as it builds on existing dual-stream architectures.
The paper tackled the problem of single image reflection removal by enhancing dual-stream architectures to better exploit inter-layer complementarity, resulting in state-of-the-art separation quality on multiple datasets with reduced computational cost and model complexity.
Although dual-stream architectures have achieved remarkable success in single image reflection removal, they fail to fully exploit inter-layer complementarity in their physical modeling and network design, which limits the quality of image separation. To address this fundamental limitation, we propose two targeted improvements to enhance dual-stream architectures: First, we introduce a novel inter-layer complementarity model where low-frequency components extracted from the residual layer interact with the transmission layer through dual-stream architecture to enhance inter-layer complementarity. Meanwhile, high-frequency components from the residual layer provide inverse modulation to both streams, improving the detail quality of the transmission layer. Second, we propose an efficient inter-layer complementarity attention mechanism which first cross-reorganizes dual streams at the channel level to obtain reorganized streams with inter-layer complementary structures, then performs attention computation on the reorganized streams to achieve better inter-layer separation, and finally restores the original stream structure for output. Experimental results demonstrate that our method achieves state-of-the-art separation quality on multiple public datasets while significantly reducing both computational cost and model complexity.