Training-free method improves finding relevant video moments for text queries
TF-PRVR: Training-Free Partially Relevant Video Retrieval
Computer Vision and Pattern Recognition
Summary
Finding specific moments related to a text search inside long videos can be tricky because current methods break videos into fixed parts and need special training that may not work well on other types of videos. The authors propose a new approach that doesn’t require any training. Instead, it uses existing video and language understanding tools to divide videos into meaningful segments and check for matching content at multiple levels. This method works consistently well across different video collections without getting stuck on details from any one set.
What this means in practice
- •For video platform developers: Improve video search features by accurately locating parts of long videos matching user text queries without additional training.
- •For media monitoring teams: Use training-free retrieval to quickly scan diverse video content streams for moments related to specified keywords or topics.
Authors
Giyeol Kim, Chanho Eom
Abstract
Partially Relevant Video Retrieval (PRVR) aims to retrieve untrimmed videos containing moments relevant to a given text query. Despite recent progress, existing PRVR methods suffer from two key limitations: a fixed video decomposition scheme that causes semantic dilution, and source-domain overfitting induced by task-specific training. In this paper, we propose TF-PRVR, the first training-free framework for PRVR. TF-PRVR leverages frozen vision-language features to construct video-specific hierarchical representations. It derives temporal semantic signals from frame-level features and applies frequency-based multi-scale analysis to identify adaptive temporal boundaries, producing hierarchical segments with coherent event-level semantics. Built on these segments, TF-PRVR constructs a unified multi-scale graph and propagates query relevance across temporally and semantically related nodes. A moment-aware scoring strategy then aggregates temporally aligned relevance across scales, emphasizing consistently supported moments while suppressing isolated false responses. Without task-specific training, TF-PRVR preserves the general-purpose alignment capability of pre-trained vision-language models and avoids dataset-specific overfitting. Extensive experiments demonstrate consistent performance across datasets with diverse visual and temporal characteristics, suggesting a practical direction for training-free PRVR.