Water Reflection Detection Using Symmetric Attention

2026-07-12

AI Summary

This work addresses the challenge posed by water surface reflections, which often cause computer vision systems to confuse objects with their mirror images, leading to false positives and missed detections in tasks such as object detection and semantic segmentation. To tackle this issue, the authors propose SAWRD-Net, a symmetry-aware water reflection detection network that uniquely integrates dihedral group-equivariant convolutions with a multi-scale symmetry attention mechanism to effectively model reflection symmetry. Furthermore, they introduce a matrix factorization-based decoder to directly regress key points defining the reflection axis. Evaluated on the largest available water reflection dataset, the method achieves a true positive rate of 0.890, significantly outperforming existing approaches and offering a novel pathway for robust visual understanding under reflective interference.

Abstract

Reflections of water pose a significant challenge for computer vision systems, as standard deep learning models frequently confuse objects with their mirror images, producing spurious false positives and negatives in tasks such as object detection and semantic segmentation. As a result, detecting reflection axes in natural-water scenes is pivotal for reliable object detection and scene understanding. To mitigate this issue, we leverage the intrinsic imperfect reflective symmetry of water and introduce a Symmetry-Aware Water Reflection Detection Network, namely, SAWRD-Net, that couples dihedral group-equivariant convolutions with a matrix-decomposition decoder in an end-to-end framework. First, dihedral group convolutional layers extract geometry-consistent feature maps that explicitly encode both rotational and mirror symmetries. A Multi-scale Reflection Equivariant block then aggregates features across scales and employs a symmetric-attention mechanism to highlight reflection-relevant regions. The proposed matrix-decomposition decoder factorizes high-dimensional features into compact low-rank parameter and confidence spaces, after which the network directly regresses keypoints on the reflection axis. Then a robust principal component analysis fits the final axis. Evaluated on the largest available water reflection scene data set, SAWRD-Net achieves a true-positive rate of 0.890 against human annotations, outperforming all existing water reflection detectors.

Problem

• water reflection
• reflection detection
• computer vision
• symmetry
• object detection

Innovation

• symmetric attention
• dihedral group-equivariant convolution
• matrix-decomposition decoder
• reflection axis detection
• water reflection