Active Learning For Finely-categorized Image-text Retrieval By Selecting Hard Negative Unpaired Samples

Abstract

Securing a sufficient amount of paired data is important to train an image-text retrieval (ITR) model, but collecting paired data is very expensive. To address this issue, in this paper, we propose an active learning algorithm for ITR that can collect paired data cost-efficiently. Previous studies assume that image-text pairs are given and their category labels are asked to the annotator. However, in the recent ITR studies, the importance of category label is decreased since a retrieval model can be trained with only image-text pairs. For this reason, we set up an active learning scenario where unpaired images (or texts) are given and the annotator provides corresponding texts (or images) to make paired data. The key idea of the proposed AL algorithm is to select unpaired images (or texts) that can be hard negative samples for existing texts (or images). To this end, we introduce a novel scoring function to choose hard negative samples. We validate the effectiveness of the proposed met

Stats

• citations1
• S2 citations—
• github stars0
• HF likes0
• heat score2.26
• arxiv keyjo2024active

Related papers

• Integrating Listwise Ranking Into Pairwise-based Image-text Retrieval (2023)9.16
• Information-theoretic Active Learning For Content-based Image Retrieval (2018)5.24
• Active Learning Via Classifier Impact And Greedy Selection For Interactive Image Retrieval (2024)0.95
• Robust Remote Sensing Image-text Retrieval With Noisy Correspondence (2026)1.24
• Lexlip: Lexicon-bottlenecked Language-image Pre-training For Large-scale Image-text Retrieval (2023)10.85
• Annotation Cost Efficient Active Learning For Content Based Image Retrieval (2023)5.84
• Anatomy-aware Conditional Image-text Retrieval (2025)0.00
• Negative Sample Is Negative In Its Own Way: Tailoring Negative Sentences For Image-text Retrieval (2021)3.81