关注微信公众号：人工智能前沿讲习班，公众号对话框回复“司晨阳”获取PPT。
基于人体骨架的行为识别是一个重要而且具有挑战性的计算机视觉任务。人体图像视频不仅包含了复杂的背景，还有光照变化、人体外貌变化等不确定因素，这使得基于图像视频的行为识别具有一定的局限性。相比图像视频，人体骨架视频可以很好地克服这些不确定因素的影响，所以基于人体骨架的行为识别受到越来越多的关注。
讲者介绍
司晨阳：中国科学院自动化研究所在读博士，本科毕业于郑州大学，已在CVPR、ECCV上发表论文，目前主要研究兴趣为行为识别方向，欢迎感兴趣的小伙伴一起交流讨论。
报告题目：Recent Advances on Skeleton-Based Action Recognition
报告摘要：基于人体骨架的行为识别是计算机视觉中的一个热点问题，相比图像视频，人体骨架视频可以很好地克服光照变化、人体外貌变化等不确定因素的影响，所以基于人体骨架的行为识别受到越来越多的关注。人体骨架序列不仅包含了时序特征，而且还包含了人体的空间结构特征，如何有效地从人体骨架序列中提取具有判别性的空间和时间特征是一个有待解决的问题。我们在这次分享会中主要介绍一下基于人体骨架的行为识别的最新进展。
Spotlight：了解基于人体骨架的行为识别的最新进展，存在难点分析。
论文推荐
Skeleton-Based Action Recognition with Spatial Reasoning and Temporal Stack Learning
推荐理由：This paper proposes a novel model with spatial reasoning and temporal stack learning for skeleton-based action recognition. The spatial reasoning network can capture the high-level spatial structural information within each frame, while the temporal stack learning network can model the detailed temporal dynamics of skeleton sequences.
View adaptive recurrent neural networks for high performance human action recognition from skeleton data
推荐理由：This paper presents an end-to-end view adaptation model for human action recognition from skeleton data. it is capable of regulating the observation viewpoints to the suitable ones by itself, with the optimization target of maximizing recognition performance.
Spatial temporal graph convolutional networks for skeleton-based action recognition
推荐理由：This paper proposes a generic graph-based formulation for modeling dynamic skeletons, which is the first that applies graph-based neural networks for this task
Co-occurrence feature learning from skeleton data for action recognition and detection with hierarchical aggregation
推荐理由：This paper proposes an end-to-end convolutional co-occurrence feature learning framework. The co-occurrence features are learned with a hierarchical methodology, in which different levels of contextual information are aggregated gradually.
An end-to-end spatio-temporal attention model for human action recognition from skeleton data
推荐理由：This paper employs a spatial-temporal attention model based on LSTM to select discriminative spatial and temporal features.
参考资料
https://www.bilibili.com/video/BV1wt411p7Ut/
https://bbs.sffai.com/d/34-recent-advances-on-skeleton-based-action-recognition