Human-in-the-loop adaptation in group activity feature learning for team sports video retrieval

2026-01-01

Computer Vision and Image Understanding

AI Summary

This work addresses the scarcity of annotated group activity data in team sports video retrieval by proposing a self-supervised feature learning framework that does not require predefined category labels. The approach integrates human-in-the-loop self-supervised pretraining with a data-efficient interactive fine-tuning mechanism, dynamically refining the feature space through user feedback and supported by an efficient video sampling strategy. Experiments on two team sports datasets demonstrate that the proposed method significantly improves retrieval performance in unsupervised settings, while ablation studies confirm the effectiveness of each component.

Technology Category

• Computer Vision: Video Understanding & Activity Analysis
• Machine Learning: Unsupervised & Self-Supervised Learning
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Problem

• human-in-the-loop
• group activity feature learning
• video retrieval
• team sports
• self-supervised learning

Innovation

• human-in-the-loop
• group activity feature learning
• self-supervised learning
• interactive fine-tuning
• contrastive learning