SFMambaNet: Spectral-Frequency Enhanced Selective State Space Model for Correspondence Pruning
2026-06-03 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors present SFMambaNet, a new method to better identify matching points between two views by improving how networks understand geometric relationships. They introduce special blocks that use frequency information to reduce noise and highlight important features, helping the network tell correct matches (inliers) from wrong ones (outliers) more accurately. Their approach combines local and global processing with frequency-based techniques, leading to improved performance on difficult tasks compared to existing methods.
Correspondence pruningGraph Neural NetworksSpectral positional encodingFrequency domainMamba networkAttention mechanismInlier-outlier separationLocal and global feature modelingFrequency gatingGeometric consistency
Authors
Zhihua Wang, Yanping Li, Yizhang Liu
Abstract
Correspondence pruning aims to identify inliers from an initial set of correspondences. Most existing Graph Neural Network (GNN)-based methods rely on geometric features mapped from coarse Euclidean coordinates, which struggle to capture the subtle geometric consistencies presented by inliers. While Mamba-based methods possess global receptive fields and long sequence modeling capabilities, they tend to accumulate substantial inconsistent features within the hidden state space, making it difficult to distinguish inliers from outliers. In this paper, we integrate frequency domain perception into this task for the first time and propose SFMambaNet, a novel Spectral-Frequency enhanced Mamba-based two-view correspondence pruning network. Our method is collaboratively composed of two components: First, we design a Local Spectral-Geometric Attention (LSGA) block. LSGA incorporates spectral positional encoding into local graph interactions and introduces multi-scale Mamba processing to enhance the capture of subtle geometric consistencies and improve local feature discriminability. Building upon this, we design a Spectral-Integrated Global Mamba (SIGM) block. SIGM embeds a frequency gating mechanism within the state space, utilizing the frequency information provided by LSGA to explicitly suppress high-frequency noise accumulation within hidden states and mitigate the propagation of inconsistent features. This enhances inlier-outlier separability and achieves robust global context modeling capabilities with nearly linear complexity. Extensive experiments demonstrate that SFMambaNet outperforms current state-of-the-art methods on several challenging tasks. The code is available at https://github.com/Kirito14IT/SFMambaNet.