arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.06403cs.CVcs.AI
收藏

利用三元组损失实现无监督动作分割

Leveraging triplet loss for unsupervised action segmentation

Institut de Robòtica i Informàtica Industrial, CSIC-UPC(西班牙高等科学研究理事会-加泰罗尼亚理工大学机器人与工业信息研究所);Amazon Alexa AI(亚马逊Alexa人工智能部门)

E. Bueno-Benito, B. Tura, M. Dimiccoli

AI总结:

本研究提出一种完全无监督的动作分割框架,基于浅层网络与带新型选择策略的三元组损失从单视频学习动作表征,在两个基准数据集上取得了有竞争力的无监督分割性能。

摘要:

本文提出了一种全新的完全无监督框架,无需任何训练数据,仅从单段输入视频本身学习适用于动作分割任务的动作表征。该方法是一种深度度量学习方案,基于浅层网络构建,采用在相似度分布上运算的三元组损失,还配备了新颖的三元组选择策略,可有效建模时间先验与语义先验,以在新的表征空间中发现动作。在此框架下,我们通过学习得到的动作表征成功恢复出时间边界,质量优于现有无监督方法。我们在两个广泛使用的动作分割任务基准数据集上评估了所提方法,对学习到的表征应用通用聚类算法后,该方法取得了具有竞争力的性能。

英文摘要:

In this paper, we propose a novel fully unsupervised framework that learns action representations suitable for the action segmentation task from the single input video itself, without requiring any training data. Our method is a deep metric learning approach rooted in a shallow network with a triplet loss operating on similarity distributions and a novel triplet selection strategy that effectively models temporal and semantic priors to discover actions in the new representational space. Under these circumstances, we successfully recover temporal boundaries in the learned action representations with higher quality compared with existing unsupervised approaches. The proposed method is evaluated on two widely used benchmark datasets for the action segmentation task and it achieves competitive performance by applying a generic clustering algorithm on the learned representations.

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。

↑