VALSE

VALSE 首页 活动通知 查看内容

VALSE 论文速览 第171期:基于骨架的带有遮挡的人体动作识别

2024-5-20 13:32| 发布者: 程一-计算所| 查看: 341| 评论: 0

摘要: 为了使得视觉与学习领域相关从业者快速及时地了解领域的最新发展动态和前沿技术进展,VALSE最新推出了《论文速览》栏目,将在每周发布一至两篇顶会顶刊论文的录制视频,对单个前沿工作进行细致讲解。本期VALSE论文速 ...

为了使得视觉与学习领域相关从业者快速及时地了解领域的最新发展动态和前沿技术进展,VALSE最新推出了《论文速览》栏目,将在每周发布一至两篇顶会顶刊论文的录制视频,对单个前沿工作进行细致讲解。本期VALSE论文速递选取了来自东南大学网络空间安全学院、东南大学计算机科学与工程学院的基于骨架的鲁棒人体行为识别的工作。该工作由东南大学网络空间安全学院硕士生陈振洁录制。


论文题目:

Occluded Skeleton-Based Human Action Recognition with Dual Inhibition Training

作者列表:

陈振洁 (东南大学),王洪松 (东南大学),桂杰 (东南大学)


B站观看网址:

https://www.bilibili.com/video/BV1WJ4m1j7vF/



论文摘要:

Recently, skeleton-based human action recognition has received widespread attention in computer vision and multimedia communities. However, most existing research focuses on improving the recognition accuracy on complete skeleton data, while ignoring the performance on the incomplete skeleton data with occlusion or noise. This paper addresses occluded and noise-robust skeleton-based action recognition and presents a novel Dual Inhibition Training strategy. Specifically, we propose Part-aware and Dual-inhibition Graph Convolutional Network (PDGCN), which comprises of three parts: Input Skeleton Inhibition (ISI), Part-Aware Representation Learning (PARL) and Predicted Score Inhibition (PSI). The ISI and PSI are plug and play modules which could encourage the model to learn discriminative features from diversified body joints by effectively simulating key body part occlusions and random occlusions. The PARL module learns both the global and local representations from the whole body and body parts, respectively, and progressively fuses them during representation learning to enhance the model robustness under occlusions. Finally, we design different settings for occluded skeleton-based human action recognition to deep study this problem and better evaluate different approaches. Our approach considerably outperforms previous state-of-the-art methods and particularly exceeds the baselines with an average improvement of 10% on the Toyota Smarthome dataset.


论文链接:

[https://dl.acm.org/doi/10.1145/3581783.3612170]

 

视频讲者简介:

陈振洁,东南大学硕士研究生在读。第三作者简介:桂杰博士,东南大学青年首席教授、博士生导师,国家青年高层次人才(国家级四青人才),目前主要从事模式识别、机器学习、图像处理、计算机视觉及人工智能安全等方面的研究,是CCF杰出会员、IEEE高级会员、 ACM高级会员、中国人工智能学会高级会员和中国图象图形学学会高级会员。荣获2021年度江苏省计算机学会“青年科技奖”。担任JCR一区期刊IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT), Artificial Intelligence Review, Neural Networks和Neurocomputing的编委(Associate editor, AE),是多个国际SCI期刊TPAMI,TNNLS, TCYB, TIP, TIFS等的审稿人。近年来,在国际学术期刊和会议上发表学术论文七十多篇,有多篇SCI高被引论文。 


个人主页:

https://guijiejie.github.io/index.html



特别鸣谢本次论文速览主要组织者:

月度轮值AC:张瑞茂 (香港中文大学 (深圳))


活动参与方式

1、VALSE每周举行的Webinar活动依托B站直播平台进行,欢迎在B站搜索VALSE_Webinar关注我们!

直播地址:

https://live.bilibili.com/22300737;

历史视频观看地址:

https://space.bilibili.com/562085182/ 


2、VALSE Webinar活动通常每周三晚上20:00进行,但偶尔会因为讲者时区问题略有调整,为方便您参加活动,请关注VALSE微信公众号:valse_wechat 或加入VALSE QQ T群,群号:863867505);


*注:申请加入VALSE QQ群时需验证姓名、单位和身份缺一不可。入群后,请实名,姓名身份单位。身份:学校及科研单位人员T;企业研发I;博士D;硕士M。


3、VALSE微信公众号一般会在每周四发布下一周Webinar报告的通知。


4您也可以通过访问VALSE主页:http://valser.org/ 直接查看Webinar活动信息。Webinar报告的PPT(经讲者允许后),会在VALSE官网每期报告通知的最下方更新。

小黑屋|手机版|Archiver|Vision And Learning SEminar

GMT+8, 2024-6-22 08:15 , Processed in 0.014645 second(s), 14 queries .

Powered by Discuz! X3.4

Copyright © 2001-2020, Tencent Cloud.

返回顶部