Video-QTR: Query-Driven Temporal Reasoning Framework for Lightweight Video Understanding

TL;DR

Video-QTR introduces a query-driven, adaptive framework for long-video understanding that reduces computational load and achieves state-of-the-art results by focusing on relevant frames based on semantic queries.

Contribution

It proposes a novel lightweight, query-guided reasoning framework that dynamically allocates perceptual resources, improving efficiency and scalability in long-video comprehension.

Findings

Reduces input frame consumption by up to 73%.

Achieves state-of-the-art performance on five benchmarks.

Demonstrates effective adaptive resource allocation based on queries.

Abstract

The rapid development of multimodal large-language models (MLLMs) has significantly expanded the scope of visual language reasoning, enabling unified systems to interpret and describe complex visual content. However, applying these models to long-video understanding remains computationally intensive. Dense frame encoding generates excessive visual tokens, leading to high memory consumption, redundant computation, and limited scalability in real-world applications. This inefficiency highlights a key limitation of the traditional process-then-reason paradigm, which analyzes visual streams exhaustively before semantic reasoning. To address this challenge, we introduce Video-QTR (Query-Driven Temporal Reasoning), a lightweight framework that redefines video comprehension as a query-guided reasoning process. Instead of encoding every frame, Video-QTR dynamically allocates perceptual resources based on the semantic intent of the query, creating an adaptive feedback loop between reasoning and perception. Extensive experiments across five benchmarks: MSVD-QA, Activity Net-QA, Movie Chat, and Video MME demonstrate that Video-QTR achieves state-of-the-art performance while reducing input frame consumption by up to 73%. These results confirm that query-driven temporal reasoning provides an efficient and scalable solution for video understanding.