Text-Video Retrieval With Global-Local Contrastive Consistency Learning

2026-05-18

AI Summary

This work addresses the challenges of semantic misalignment between text and video modalities and the high computational cost of existing attention mechanisms by proposing a lightweight global–local contrastive consistency learning framework. The method leverages text guidance to generate semantically aligned frame-level and video-level features and introduces a parameter-free global–local interaction module that enhances multi-granularity semantic associations without adding extra model parameters. Additionally, a contrastive score consistency loss is incorporated to strengthen the model’s ability to discriminate between positive and negative samples. Extensive experiments on three benchmark datasets—MSR-VTT, DiDeMo, and VATEX—demonstrate that the proposed approach achieves superior text-to-video retrieval performance while significantly reducing computational overhead compared to state-of-the-art methods.

Abstract

Text-video retrieval aims to find the most semantically similar videos with given text queries. However, since videos contain more diverse content than texts, the main semantics expressed by each text-video pair is often partially relevant. The primary methods involve the utilization of language-video attention module to align texts and videos. Though effective, this paradigm inevitably introduces prohibitive computational overhead, resulting in inefficient retrieval. In this paper, we propose a simple yet effective method called Global-Local Contrastive Consistency Learning (GLCCL) to achieve texts and videos semantics alignment. Specifically, we design a parameter-free Global-Local Interaction Module (GLIM) to generate semantic-related frame and video features in a text-guided manner. Furthermore, a Contrastive Score Consistency (CSC) loss is developed to promote consistency learning among different scores on positive pairs and suppress consistency learning on negative pairs. Empirical evidence suggests that CSC loss provides the model with robust discriminative power between positives and hard negatives. Extensive experiments on three benchmark datasets, including MSR-VTT, DiDeMo and VATEX, demonstrate the effectiveness and superiority of our approach.

Problem

• text-video retrieval
• semantic alignment
• computational overhead
• partial relevance
• inefficient retrieval

Innovation

• Global-Local Contrastive Consistency Learning
• Parameter-free Interaction Module
• Contrastive Score Consistency
• Text-Video Retrieval
• Semantic Alignment