Noisy Correspondence

Noisy correspondence, the presence of mismatched data pairs in cross-modal datasets, hinders accurate cross-modal retrieval and matching. Current research focuses on developing robust methods to identify and mitigate these mismatches, employing techniques like geometrical structure consistency learning, dual attention mechanisms, and memory banks to improve the reliability of learned correspondences. These advancements are crucial for improving the performance of various applications, including image-text retrieval, point cloud matching, and person re-identification, where noisy data is prevalent and impacts the accuracy of learned models. The ultimate goal is to enable reliable cross-modal learning even with imperfect or automatically collected data.

Papers

May 27, 2024

Mitigating Noisy Correspondence by Geometrical Structure Consistency Learning
Zihua Zhao, Mengxi Chen, Tianjie Dai, Jiangchao Yao, Bo han, Ya Zhang, Yanfeng Wang
Cross Modal Geometric Consistency Noisy Correspondence

March 28, 2024

PointCloud-Text Matching: Benchmark Datasets and a Baseline
Yanglin Feng, Yang Qin, Dezhong Peng, Hongyuan Zhu, Xi Peng, Peng Hu
Point Cloud Baseline Result Benchmark Datasets Point Cloud Data Cross Modal Matching Noisy Correspondence

March 13, 2024

REPAIR: Rank Correlation and Noisy Pair Half-replacing with Memory for Noisy Correspondence
Ruochen Zheng, Jiahao Hong, Changxin Gao, Nong Sang
Cross Modal Memory Trace Soft Label Third Position Repair Noisy Pair Cross Modal Matching Noisy Correspondence Rank Correlation Noisy Correspondence Learning

December 27, 2023

Noisy Correspondence Learning with Self-Reinforcing Errors Mitigation
Zhuohang Dang, Minnan Luo, Chengyou Jia, Guang Dai, Xiaojun Chang, Jingdong Wang
Cross Modal Retrieval Error Mitigation Noisy Correspondence Noisy Correspondence Learning

December 10, 2023

Negative Pre-aware for Noisy Cross-modal Matching
Xu Zhang, Hao Li, Mang Ye
Vision Language Cross Modal Noise Robust Learning Cross Modal Matching Noisy Correspondence

October 26, 2023

Cross-modal Active Complementary Learning with Self-refining Correspondence
Yang Qin, Yuan Sun, Dezhong Peng, Joey Tianyi Zhou, Xi Peng, Peng Hu
Cross Modal Self Supervised Correspondence Noisy Correspondence Image Text Benchmark

August 19, 2023

Noisy-Correspondence Learning for Text-to-Image Person Re-identification
Yang Qin, Yingke Chen, Dezhong Peng, Xi Peng, Joey Tianyi Zhou, Peng Hu
Image Text Pair Noisy Correspondence Noisy Correspondence Learning

June 5, 2023

A2B: Anchor to Barycentric Coordinate for Robust Correspondence
Weiyue Zhao, Hao Lu, Zhiguo Cao, Xin Li
Visual Information Anchor Learning Correspondence Noisy Correspondence Image Correspondence Barycentric Coordinate Sparse Correspondence Coordinate Representation

May 16, 2023

Learning Correspondence Uncertainty via Differentiable Nonlinear Least Squares
Dominik Muhle, Lukas Koestler, Krishna Murthy Jatavallabhula, Daniel Cremers
Least Square Relative Pose Product Specific Position Information Closed Form Differentiable Expression Learning Correspondence Epipolar Constraint Noisy Correspondence

April 13, 2023

Noisy Correspondence Learning with Meta Similarity Correction
Haochen Han, Kaiyao Miao, Qinghua Zheng, Minnan Luo
Multimodal Learning Cross Modal Retrieval Noisy Correspondence Noisy Correspondence Learning Meta Similarity Correction Network Meta Label Correction

December 8, 2022

Graph Matching with Bi-level Noisy Correspondence
Yijie Lin, Mouxing Yang, Jun Yu, Peng Hu, Changqing Zhang, Xi Peng
Graph Matching Noisy Correspondence

November 29, 2021

Multi-instance Point Cloud Registration by Efficient Correspondence Clustering
Weixuan Tang, Danping Zou
Point Cloud Global Outlier Noisy Correspondence Distance Matrix Correspondence Prediction Multi Instance