arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2406.03447cs.CVcs.AIcs.LG
收藏

FILS:在语义语言空间中的自监督视频特征预测

FILS: Self-Supervised Video Feature Prediction In Semantic Language Space

University of Surrey(萨里大学)

Mona Ahmadian, Frank Guerin, Andrew Gilbert

AI总结:

本文提出FILS,一种在语义语言空间中进行自监督视频特征预测的新方法,通过逐块视频-文本对比策略将视觉特征转换至语言空间并作为掩码预测目标,在多个下游动作识别数据集上实现了最先进的性能且计算效率更高。

摘要:

本文展示了一种用于学习语义视频表示的自监督方法。近期的视觉研究表明,视觉掩码策略和自然语言监督有助于开发可迁移的视觉预训练。我们的目标是通过在预训练过程中以完全自监督的方式利用与视频内容相关的文本,来实现更具语义的视频表示。为此,我们提出了FILS,一种新颖的自监督视频特征预测方法(在语义语言空间中进行视频特征预测,FILS)。视觉模型通过在语言空间中正确预测掩码特征的语义,能够捕获有价值的结构化信息。该方法使用逐块视频-文本对比策略进行学习,其中文本表示作为原型,将视觉特征转换到语言空间,随后将其作为目标,利用我们的掩码编码器-解码器结构进行具有语义意义的特征预测。FILS在下游动作识别任务中展现出卓越的迁移能力,使用ViT-Base在Epic-Kitchens、Something-SomethingV2、Charades-Ego和EGTEA等具有挑战性的以自我为中心的数据集上实现了最先进的性能。与以往工作相比,我们高效的方法需要更少的计算量和更小的批次。

英文摘要:

This paper demonstrates a self-supervised approach for learning semantic video representations. Recent vision studies show that a masking strategy for vision and natural language supervision has contributed to developing transferable visual pretraining. Our goal is to achieve a more semantic video representation by leveraging the text related to the video content during the pretraining in a fully self-supervised manner. To this end, we present FILS, a novel self-supervised video Feature prediction In semantic Language Space (FILS). The vision model can capture valuable structured information by correctly predicting masked feature semantics in language space. It is learned using a patch-wise video-text contrastive strategy, in which the text representations act as prototypes for transforming vision features into a language space, which are then used as targets for semantically meaningful feature prediction using our masked encoder-decoder structure. FILS demonstrates remarkable transferability on downstream action recognition tasks, achieving state-of-the-art on challenging egocentric datasets, like Epic-Kitchens, Something-SomethingV2, Charades-Ego, and EGTEA, using ViT-Base. Our efficient method requires less computation and smaller batches compared to previous works.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑