Saved in:
Bibliographic Details
Main Authors: Yin, Yufei, Xing, Yuchen, Meng, Qianke, Chen, Minghao, Yang, Yan, Yu, Zhou
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.02891
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910100079247360
author Yin, Yufei
Xing, Yuchen
Meng, Qianke
Chen, Minghao
Yang, Yan
Yu, Zhou
author_facet Yin, Yufei
Xing, Yuchen
Meng, Qianke
Chen, Minghao
Yang, Yan
Yu, Zhou
contents Understanding long videos requires extracting query-relevant information from long sequences under tight compute budgets. Existing text-then-LLM pipelines lose fine-grained visual cues, while video-based multimodal large language models (MLLMs) can keep visual details but are too frame-hungry and computationally expensive. In this work, we aim to harness MLLMs for efficient video understanding. We propose ProVCA, a progressive video condensation agent that iteratively locates key video frames at multiple granularities. ProVCA first adopts a segment localization module to identify the video segment relevant to the query, then a snippet selection module to select important snippets based on similarity, and finally a keyframe refinement module to pinpoint specific keyframes in those snippets. By progressively narrowing the scope from coarse segments to fine frames, ProVCA identifies a small set of keyframes for MLLM-based reasoning. ProVCA achieves state-of-the-art zero-shot accuracies of 69.3\% on EgoSchema, 80.5\% on NExT-QA, and 77.7\% on IntentQA, while using fewer frames than previous training-free methods.
format Preprint
id arxiv_https___arxiv_org_abs_2604_02891
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Progressive Video Condensation with MLLM Agent for Long-form Video Understanding
Yin, Yufei
Xing, Yuchen
Meng, Qianke
Chen, Minghao
Yang, Yan
Yu, Zhou
Computer Vision and Pattern Recognition
Understanding long videos requires extracting query-relevant information from long sequences under tight compute budgets. Existing text-then-LLM pipelines lose fine-grained visual cues, while video-based multimodal large language models (MLLMs) can keep visual details but are too frame-hungry and computationally expensive. In this work, we aim to harness MLLMs for efficient video understanding. We propose ProVCA, a progressive video condensation agent that iteratively locates key video frames at multiple granularities. ProVCA first adopts a segment localization module to identify the video segment relevant to the query, then a snippet selection module to select important snippets based on similarity, and finally a keyframe refinement module to pinpoint specific keyframes in those snippets. By progressively narrowing the scope from coarse segments to fine frames, ProVCA identifies a small set of keyframes for MLLM-based reasoning. ProVCA achieves state-of-the-art zero-shot accuracies of 69.3\% on EgoSchema, 80.5\% on NExT-QA, and 77.7\% on IntentQA, while using fewer frames than previous training-free methods.
title Progressive Video Condensation with MLLM Agent for Long-form Video Understanding
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.02891