ARD-REFSM: Enhancing Reflection Symmetry Detection with Asymmetric Denoising and Rotation Equivariance

2026-07-30

AI Summary

This work addresses the challenges in reflection symmetry detection caused by asymmetric distractions and arbitrary-oriented symmetry patterns, which degrade performance due to the lack of rotational equivariance in conventional convolutional networks, leading to unstable features. To overcome this, the authors propose a dual-module collaborative framework: an Asymmetric Region Denoising (ARD) module suppresses background interference, while a Rotation-Equivariant Feature Similarity Matching (REFSM) module enhances symmetry axis prediction accuracy by enforcing feature consistency between original and rotated images. This study is the first to integrate asymmetric denoising with rotation-equivariant matching in a unified architecture and introduces GMSYM, a more challenging benchmark dataset. The method achieves state-of-the-art performance across five datasets—DENDI, NYU, LDRS, SDRW, and GMSYM—demonstrating significantly improved accuracy and robustness.

Abstract

Reflection symmetry detection remains challenging due to interference from asymmetric regions and arbitrary orientations of symmetric patterns. Asymmetric regions introduce background clutter that disrupts symmetric pattern matching, whereas conventional convolutional neural networks lack rotation equivariance, leading to inconsistent feature representations under rotational transformations. To address these issues, we propose an Asymmetric Region Denoising (ARD) module and a Rotation Equivariant Feature Similarity Matching (REFSM) module. The ARD module suppresses asymmetric interference to refine symmetric patterns, while the REFSM module enhances rotation equivariance through feature similarity matching between original and rotated images. Specifically, our dual-input REFSM framework leverages rotation loss to maximize consistency between the score maps of original and rotated images, thereby enabling precise prediction of rotation-equivariant symmetry axes. Furthermore, we introduce GMSYM, a new benchmark dataset that categorizes images into diverse scenarios and incorporates various interferences to address the limitations of existing reflection symmetry detection benchmarks. Extensive experiments on four standard datasets (DENDI, NYU, LDRS, SDRW) and our proposed GMSYM dataset demonstrate that our method achieves state-of-the-art performance in both accuracy and robustness.

Problem

• reflection symmetry detection
• asymmetric regions
• rotation equivariance
• background clutter
• symmetric pattern matching

Innovation

• Asymmetric Region Denoising
• Rotation Equivariance
• Reflection Symmetry Detection
• Feature Similarity Matching
• GMSYM Dataset