Structured Video-Language Modeling with Temporal Grouping and Spatial Grounding

TL;DR

This paper introduces S-ViLM, a novel video-language model that enhances fine-grained spatial and temporal understanding through inter-clip spatial grounding and intra-clip temporal grouping, improving performance on multiple downstream tasks.

Contribution

The paper proposes a new framework, S-ViLM, that captures region-object correspondences and scene changes by exploiting intrinsic structures of video and text modalities.

Findings

Outperforms existing methods on four downstream tasks

Achieves significant improvements in text-video retrieval and video question answering

Enhances temporal localization and semantic reasoning capabilities

Abstract

Existing video-language pre-training methods primarily focus on instance-level alignment between video clips and captions via global contrastive learning but neglect rich fine-grained local information in both videos and text, which is of importance to downstream tasks requiring temporal localization and semantic reasoning. A powerful model is expected to be capable of capturing region-object correspondences and recognizing scene changes in a video clip, reflecting spatial and temporal granularity, respectively. To strengthen model's understanding into such fine-grained details, we propose a simple yet effective video-language modeling framework, S-ViLM, by exploiting the intrinsic structures of these two modalities. It includes two novel designs, inter-clip spatial grounding and intra-clip temporal grouping, to promote learning region-object alignment and temporal-aware features, simultaneously. Comprehensive evaluations demonstrate that S-ViLM performs favorably against existing approaches in learning more expressive representations. Specifically, S-ViLM surpasses the state-of-the-art methods substantially on four representative downstream tasks, covering text-video retrieval, video question answering, video action recognition, and temporal action localization.