作者：
Xin ZhangHongzhi FengM. Shamim HossainYinzhuo ChenHongbo WangYuyu YinHangzhou Dianzi University
China Key Laboratory of Complex Systems Modeling and Simulation Ministry of Education China Zhoushan Tongbo Marine Electronic Information Research Institute Hangzhou Dianzi University China and Yunnan Key Laboratory of Service Computing Yunnan University of Finance and Economics China Hangzhou Dianzi University
China Department of Software Engineering
College of Computer and Information Sciences King Saud University Saudi Arabia Hangzhou Dianzi University
China Key Laboratory of Complex Systems Modeling and Simulation Ministry of Education China and Zhoushan Tongbo Marine Electronic Information Research Institute Hangzhou Dianzi University China
Action Quality Assessment (AQA) has become crucial in video analysis, finding wide applications in various domains, such as healthcare and sports. A significant challenge faced by AQA is the background bias due to the...
详细信息
Action Quality Assessment (AQA) has become crucial in video analysis, finding wide applications in various domains, such as healthcare and sports. A significant challenge faced by AQA is the background bias due to the dominance of the background in videos. Especially, the background bias tends to overshadow subtle foreground differences, which is crucial for precise action evaluation. To address the background bias issue, we propose a novel data augmentation method named Scaled Background Swap. Firstly, the background regions between different video samples are swapped to guide models focus toward the dynamic foreground regions and mitigate its sensitivity to the background during training. Secondly, the video’s foreground region is up-scaled to further enhance models’ attention to the critical foreground action information for AQA tasks. In particular, the proposed Scaled Background Swap method can effectively improve models’ accuracy and generalization by prioritizing foreground motion and swapping backgrounds. It can be flexibly applied with various video analysis models. Extensive experiments on AQA benchmarks demonstrate that Scaled Background Swap method achieves better performance than baselines. Specifically, the Spearman’s rank correlation on datasets AQA-7 and MTL-AQA reaches 0.8870 and 0.9526, respectively. The code is available at: https://***/Emy-cv/Scaled-Background Swap.
This book grew out of the First Symposium on the Personal Web, co-located with CASCON 2010 in Markham, Ontario, Canada. The purpose of the symposium was to bring together prominent researchers and practitioners from a...
详细信息
ISBN:
(数字)9783642399954
ISBN:
(纸本)9783642399947
This book grew out of the First Symposium on the Personal Web, co-located with CASCON 2010 in Markham, Ontario, Canada. The purpose of the symposium was to bring together prominent researchers and practitioners from a diverse range of research areas relevant to the advancement of science and practice relating to the Personal Web. Research on the Personal Web is an outgrowth of the Smart Internet initiative, which seeks to extend and transform the web to be centred on the user, with the web as a calm platform ubiquitously providing cognitive support to its user and his or her tasks. As with the preceding SITCON workshop (held at CASCON 2009), this symposium involved a multi-disciplinary effort that brought together researchers and practitioners in data integration; web services modelling and architecture; human-computer interaction; predictive analytics; cloud infrastructure; semantics and ontology; and industrial application domains such as health care and finance.
The discussions during the symposium dealt with different aspects of the architecture and functionality needed to make the Personal Web a reality. After the symposium the authors reworked their presentations into draft chapters that were submitted for peer evaluation and review. Every chapter went through two rounds of reviewing by at least two independent expert reviewers, and accepted chapters were then revised and are presented in this book.
暂无评论