Liu, Jiayang, et al. “Noise Mitigation for Unsupervised Cross-Domain Image Retrieval.” 2025 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2025.
Paper Link: ICME 2025
Introduction
This paper addresses noise mitigation in unsupervised cross-domain image retrieval. The method focuses on handling noisy data and improving retrieval robustness.
Key Contributions
- Noise mitigation framework for cross-domain retrieval
- Robust feature learning under noisy conditions
- Improved retrieval performance with noisy data
Methodology
The proposed approach incorporates noise mitigation strategies to learn robust features that are less sensitive to noise in cross-domain retrieval scenarios.
Figures
Note: This paper was published at ICME 2025. Images will be added once the paper is available on arXiv or from the conference proceedings.
Results
The method shows improved robustness and performance in cross-domain image retrieval when dealing with noisy data.