M-adapter: Multi-level image-to-video adaptation for video action recognition

IF 4.3 3区 计算机科学 Q2 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Rongchang Li , Tianyang Xu , Xiao-Jun Wu , Linze Li , Xiao Yang , Zhongwei Shen , Josef Kittler
{"title":"M-adapter: Multi-level image-to-video adaptation for video action recognition","authors":"Rongchang Li ,&nbsp;Tianyang Xu ,&nbsp;Xiao-Jun Wu ,&nbsp;Linze Li ,&nbsp;Xiao Yang ,&nbsp;Zhongwei Shen ,&nbsp;Josef Kittler","doi":"10.1016/j.cviu.2024.104150","DOIUrl":null,"url":null,"abstract":"<div><div>With the growing size of visual foundation models, training video models from scratch has become costly and challenging. Recent attempts focus on transferring frozen pre-trained Image Models (PIMs) to video fields by tuning inserted learnable parameters such as adapters and prompts. However, these methods require saving PIM activations for gradient calculations, leading to limited savings of GPU memory. In this paper, we propose a novel parallel branch that adapts the multi-level outputs of the frozen PIM for action recognition. It avoids passing gradients through the PIMs, thus naturally owning much lower GPU memory footprints. The proposed adaptation branch consists of hierarchically combined multi-level output adapters (M-adapters), comprising a fusion module and a temporal module. This design digests the existing discrepancies between the pre-training task and the target task with lower training costs. We show that when using larger models or on scenarios with higher demands for temporal modelling, the proposed method performs better than those with the full-parameter tuning manner. Finally, despite only tuning fewer parameters, our method achieves superior or comparable performance against current state-of-the-art methods.</div></div>","PeriodicalId":50633,"journal":{"name":"Computer Vision and Image Understanding","volume":null,"pages":null},"PeriodicalIF":4.3000,"publicationDate":"2024-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computer Vision and Image Understanding","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1077314224002315","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

Abstract

With the growing size of visual foundation models, training video models from scratch has become costly and challenging. Recent attempts focus on transferring frozen pre-trained Image Models (PIMs) to video fields by tuning inserted learnable parameters such as adapters and prompts. However, these methods require saving PIM activations for gradient calculations, leading to limited savings of GPU memory. In this paper, we propose a novel parallel branch that adapts the multi-level outputs of the frozen PIM for action recognition. It avoids passing gradients through the PIMs, thus naturally owning much lower GPU memory footprints. The proposed adaptation branch consists of hierarchically combined multi-level output adapters (M-adapters), comprising a fusion module and a temporal module. This design digests the existing discrepancies between the pre-training task and the target task with lower training costs. We show that when using larger models or on scenarios with higher demands for temporal modelling, the proposed method performs better than those with the full-parameter tuning manner. Finally, despite only tuning fewer parameters, our method achieves superior or comparable performance against current state-of-the-art methods.
M-adapter:用于视频动作识别的多级图像视频适配器
随着视觉基础模型的规模不断扩大,从头开始训练视频模型变得既昂贵又具有挑战性。最近的尝试侧重于通过调整插入的可学习参数(如适配器和提示),将冻结的预训练图像模型(PIM)转移到视频领域。然而,这些方法需要保存 PIM 激活以进行梯度计算,因此只能有限地节省 GPU 内存。在本文中,我们提出了一个新颖的并行分支,它可以调整冻结 PIM 的多级输出,用于动作识别。它避免了通过 PIM 传递梯度,从而自然而然地大大降低了 GPU 内存占用。拟议的适配分支由分层组合的多级输出适配器(M-adapters)组成,包括一个融合模块和一个时序模块。这种设计以较低的训练成本消化了预训练任务与目标任务之间的现有差异。我们的研究表明,在使用较大的模型或对时间建模要求较高的场景时,所提出的方法比采用全参数调整方式的方法表现更好。最后,尽管只调整了较少的参数,我们的方法仍然取得了优于或媲美当前最先进方法的性能。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Computer Vision and Image Understanding
Computer Vision and Image Understanding 工程技术-工程:电子与电气
CiteScore
7.80
自引率
4.40%
发文量
112
审稿时长
79 days
期刊介绍: The central focus of this journal is the computer analysis of pictorial information. Computer Vision and Image Understanding publishes papers covering all aspects of image analysis from the low-level, iconic processes of early vision to the high-level, symbolic processes of recognition and interpretation. A wide range of topics in the image understanding area is covered, including papers offering insights that differ from predominant views. Research Areas Include: • Theory • Early vision • Data structures and representations • Shape • Range • Motion • Matching and recognition • Architecture and languages • Vision systems
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信