VideoRefer Suite: Advancing Spatial-Temporal Object Understanding with Video LLM

TL;DR

The paper introduces VideoRefer Suite, a comprehensive framework including a new dataset, model, and benchmark, to enhance spatial-temporal object understanding in Video LLMs, addressing previous limitations in fine-grained video comprehension.

Contribution

It presents a new high-quality object-level video instruction dataset, a versatile spatial-temporal object encoder model, and a benchmark for detailed video understanding evaluation.

Findings

VideoRefer model achieves promising results on referring benchmarks.

The dataset VideoRefer-700K enables fine-grained spatial-temporal understanding.

The benchmark facilitates comprehensive assessment of Video LLM capabilities.

Abstract

Video Large Language Models (Video LLMs) have recently exhibited remarkable capabilities in general video understanding. However, they mainly focus on holistic comprehension and struggle with capturing fine-grained spatial and temporal details. Besides, the lack of high-quality object-level video instruction data and a comprehensive benchmark further hinders their advancements. To tackle these challenges, we introduce the VideoRefer Suite to empower Video LLM for finer-level spatial-temporal video understanding, i.e., enabling perception and reasoning on any objects throughout the video. Specially, we thoroughly develop VideoRefer Suite across three essential aspects: dataset, model, and benchmark. Firstly, we introduce a multi-agent data engine to meticulously curate a large-scale, high-quality object-level video instruction dataset, termed VideoRefer-700K. Next, we present the VideoRefer model, which equips a versatile spatial-temporal object encoder to capture precise regional and sequential representations. Finally, we meticulously create a VideoRefer-Bench to comprehensively assess the spatial-temporal understanding capability of a Video LLM, evaluating it across various aspects. Extensive experiments and analyses demonstrate that our VideoRefer model not only achieves promising performance on video referring benchmarks but also facilitates general video understanding capabilities.