Lune

CVPR Findings2026

FusionBridge: An Efficient Fusion Via Feature Disentanglement for Multi-Modal Object Re-Identification

Yali Li, Qianru Han, Xinwei He, Zhi Liu, Jinhai Xiang

Abstract

Transformer has attracted increasing interest in multimodal object re-identification (ReID), owing to its strong ability to model long-range dependencies and integrate cross-modal information. Existing Transformer-based approaches usually focus on either cross-modal alignment and fusion or modality-specific enhancement. However, how to effectively extract cross-modal semantic consistency and modality-specific fine-grained discriminative features remains a challenging problem. To this end, we develop a novel Fusion-Bridge framework that jointly leverages frequency-domain modeling and feature disentanglement. To be specific, we first design a Hybrid Feature Aggregation (HFA) module, which captures high-frequency discriminative regions in the frequency domain to enhance fine-grained discriminative ability. Then, we introduce a Shared–Private Bottleneck Fusion (SPBF) module that models cross-modal interactions through feature disentanglement and an attention bottleneck fusion mechanism, thereby aggregating common representations across modalities while preserving modality-unique ones. Extensive experiments conducted on three benchmark datasets — RGBNT201, RGBNT100, and MSVR310 — demonstrate the effectiveness and superiority of our proposed method.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Swin Transformer: Hierarchical Vision Transformer using Shifted Windows cites Learning Transferable Visual Models From Natural Language Supervision
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on17

  1. Learning Transferable Visual Models From Natural Language SupervisionAlec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh et al.ICML 2021 · 47,906 citations
  2. Swin Transformer: Hierarchical Vision Transformer using Shifted WindowsZe Liu, Yutong Lin, Yue Cao, Han Hu et al.ICCV 2021 · 31,683 citations
  3. Do Vision Transformers See Like Convolutional Neural Networks?Maithra Raghu, Thomas Unterthiner, Simon Kornblith, Chiyuan Zhang et al.NeurIPS 2021 · 1,553 citations
  4. Perceiver: General Perception with Iterative AttentionAndrew Jaegle, Felix Gimeno, Andy Brock, Oriol Vinyals et al.ICML 2021 · 1,399 citations
  5. TransReID: Transformer-based Object Re-IdentificationShuting He, Hao Luo, Pichao Wang, Fan Wang et al.ICCV 2021 · 1,172 citations

Related papers

  1. FUSE: Frequency-domain Unification and Spectral Energy Alignment for Multi-modal Object Re-IdentificationXuanhao Qi, Tom Luan, Yukang Zhang, Jinkai Zheng et al.ICML 2026
  2. Magic Tokens: Select Diverse Tokens for Multi-modal Object Re-IdentificationPingping Zhang, Yuhao Wang, Yang Liu, Zhengzheng Tu et al.CVPR 2024 · 42 citations
  3. Multi-Modal Object Re-identification via Sparse Mixture-of-ExpertsYingying Feng, Jie Li, Chi Xie, Lei Tan et al.ICML 2025
  4. Signal: Selective Interaction and Global-local Alignment for Multi-Modal Object Re-IdentificationYangyang Liu, Yuhao Wang, Pingping ZhangAAAI 2026
  5. Learning Progressive Modality-Shared Transformers for Effective Visible-Infrared Person Re-identificationHu Lu, Xuezhang Zou, Pingping ZhangAAAI 2023 · 183 citations
Dusk over the sea between two cliffs drawn in fine vertical lines