FlashVTG: Feature Layering and Adaptive Score Handling Network for Video Temporal Grounding

TL;DR

FlashVTG introduces a novel framework with feature layering and adaptive score refinement to enhance video temporal grounding, significantly improving accuracy in moment retrieval and highlight detection without additional training complexity.

Contribution

The paper proposes FlashVTG, a new method combining temporal feature layering and adaptive score handling to better capture video content variations and context, achieving state-of-the-art results.

Findings

Achieves 5.8% mAP improvement on QVHighlights for moment retrieval.

Increases mAP to 125% of previous SOTA for short-moment retrieval.

Demonstrates effectiveness across four datasets in both subtasks.

Abstract

Text-guided Video Temporal Grounding (VTG) aims to localize relevant segments in untrimmed videos based on textual descriptions, encompassing two subtasks: Moment Retrieval (MR) and Highlight Detection (HD). Although previous typical methods have achieved commendable results, it is still challenging to retrieve short video moments. This is primarily due to the reliance on sparse and limited decoder queries, which significantly constrain the accuracy of predictions. Furthermore, suboptimal outcomes often arise because previous methods rank predictions based on isolated predictions, neglecting the broader video context. To tackle these issues, we introduce FlashVTG, a framework featuring a Temporal Feature Layering (TFL) module and an Adaptive Score Refinement (ASR) module. The TFL module replaces the traditional decoder structure to capture nuanced video content variations across multiple temporal scales, while the ASR module improves prediction ranking by integrating context from adjacent moments and multi-temporal-scale features. Extensive experiments demonstrate that FlashVTG achieves state-of-the-art performance on four widely adopted datasets in both MR and HD. Specifically, on the QVHighlights dataset, it boosts mAP by 5.8% for MR and 3.3% for HD. For short-moment retrieval, FlashVTG increases mAP to 125% of previous SOTA performance. All these improvements are made without adding training burdens, underscoring its effectiveness. Our code is available at https://github.com/Zhuo-Cao/FlashVTG.