arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2003.08275cs.CV
收藏

PIC:识别长程活动的排列不变卷积

PIC: Permutation Invariant Convolution for Recognizing Long-range Activities

University of Amsterdam(阿姆斯特丹大学)

Noureldien Hussein, Efstratios Gavves, Arnold W. M. Smeulders

摘要:

卷积、自我注意和向量聚合等神经操作是识别短程动作的首选。然而,它们在建模远程活动时有三个局限性。本文提出了PIC,置换不变卷积,一种新的神经层建模的时间结构的长期活动。它有三个可取的属性。I.与标准卷积不同,PIC对其感受野内的特征的时间排列是不变的,使其有资格对弱时间结构进行建模。二.与向量聚合不同,PIC尊重本地连接性,使其能够使用级联层学习长距离时间抽象。三.与自我注意相比,PIC使用共享权重,使其更能够在长时间和嘈杂的视频中检测最具鉴别力的视觉证据。我们研究了PIC的三个属性,并证明了它在识别Charades,Breakfast和MultiThumos的长距离活动中的有效性。

英文摘要:

Neural operations as convolutions, self-attention, and vector aggregation are the go-to choices for recognizing short-range actions. However, they have three limitations in modeling long-range activities. This paper presents PIC, Permutation Invariant Convolution, a novel neural layer to model the temporal structure of long-range activities. It has three desirable properties. i. Unlike standard convolution, PIC is invariant to the temporal permutations of features within its receptive field, qualifying it to model the weak temporal structures. ii. Different from vector aggregation, PIC respects local connectivity, enabling it to learn long-range temporal abstractions using cascaded layers. iii. In contrast to self-attention, PIC uses shared weights, making it more capable of detecting the most discriminant visual evidence across long and noisy videos. We study the three properties of PIC and demonstrate its effectiveness in recognizing the long-range activities of Charades, Breakfast, and MultiThumos.

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。

↑