LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shen, Xiaoqian, Xiong, Yunyang, Zhao, Changsheng, Wu, Lemeng, Chen, Jun, Zhu, Chenchen, Liu, Zechun, Xiao, Fanyi, Varadarajan, Balakrishnan, Bordes, Florian, Liu, Zhuang, Xu, Hu, Kim, Hyunwoo J., Soran, Bilge, Krishnamoorthi, Raghuraman, Elhoseiny, Mohamed, Chandra, Vikas
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913560806817792
author Shen, Xiaoqian
Xiong, Yunyang
Zhao, Changsheng
Wu, Lemeng
Chen, Jun
Zhu, Chenchen
Liu, Zechun
Xiao, Fanyi
Varadarajan, Balakrishnan
Bordes, Florian
Liu, Zhuang
Xu, Hu
Kim, Hyunwoo J.
Soran, Bilge
Krishnamoorthi, Raghuraman
Elhoseiny, Mohamed
Chandra, Vikas
author_facet Shen, Xiaoqian
Xiong, Yunyang
Zhao, Changsheng
Wu, Lemeng
Chen, Jun
Zhu, Chenchen
Liu, Zechun
Xiao, Fanyi
Varadarajan, Balakrishnan
Bordes, Florian
Liu, Zhuang
Xu, Hu
Kim, Hyunwoo J.
Soran, Bilge
Krishnamoorthi, Raghuraman
Elhoseiny, Mohamed
Chandra, Vikas
contents Multimodal Large Language Models (MLLMs) have shown promising progress in understanding and analyzing video content. However, processing long videos remains a significant challenge constrained by LLM's context size. To address this limitation, we propose LongVU, a spatiotemporal adaptive compression mechanism thats reduces the number of video tokens while preserving visual details of long videos. Our idea is based on leveraging cross-modal query and inter-frame dependencies to adaptively reduce temporal and spatial redundancy in videos. Specifically, we leverage DINOv2 features to remove redundant frames that exhibit high similarity. Then we utilize text-guided cross-modal query for selective frame feature reduction. Further, we perform spatial token reduction across frames based on their temporal dependencies. Our adaptive compression strategy effectively processes a large number of frames with little visual information loss within given context length. Our LongVU consistently surpass existing methods across a variety of video understanding benchmarks, especially on hour-long video understanding tasks such as VideoMME and MLVU. Given a light-weight LLM, our LongVU also scales effectively into a smaller size with state-of-the-art video understanding performance.
format Preprint
id arxiv_https___arxiv_org_abs_2410_17434
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding
Shen, Xiaoqian
Xiong, Yunyang
Zhao, Changsheng
Wu, Lemeng
Chen, Jun
Zhu, Chenchen
Liu, Zechun
Xiao, Fanyi
Varadarajan, Balakrishnan
Bordes, Florian
Liu, Zhuang
Xu, Hu
Kim, Hyunwoo J.
Soran, Bilge
Krishnamoorthi, Raghuraman
Elhoseiny, Mohamed
Chandra, Vikas
Computer Vision and Pattern Recognition
Multimodal Large Language Models (MLLMs) have shown promising progress in understanding and analyzing video content. However, processing long videos remains a significant challenge constrained by LLM's context size. To address this limitation, we propose LongVU, a spatiotemporal adaptive compression mechanism thats reduces the number of video tokens while preserving visual details of long videos. Our idea is based on leveraging cross-modal query and inter-frame dependencies to adaptively reduce temporal and spatial redundancy in videos. Specifically, we leverage DINOv2 features to remove redundant frames that exhibit high similarity. Then we utilize text-guided cross-modal query for selective frame feature reduction. Further, we perform spatial token reduction across frames based on their temporal dependencies. Our adaptive compression strategy effectively processes a large number of frames with little visual information loss within given context length. Our LongVU consistently surpass existing methods across a variety of video understanding benchmarks, especially on hour-long video understanding tasks such as VideoMME and MLVU. Given a light-weight LLM, our LongVU also scales effectively into a smaller size with state-of-the-art video understanding performance.
title LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.17434