作者机构：
[Chengxiu Li] School of Physical Education, University of South China, Hengyang, 421001, Hunan, PR China;[Ni Duan] Red Sun Experimental School, Hengyang, 421001, Hunan, PR China
通讯机构：
[Chengxiu Li] S;School of Physical Education, University of South China, Hengyang, 421001, Hunan, PR China
关键词：
Human pose estimation;Lightweight human pose estimator;Physical education application
摘要：
Current human pose estimation models adopt heavy backbones and complex feature enhance- ment modules to pursue higher accuracy. However, they ignore the need for model efficiency in real-world applications. In real-world scenarios such as sports teaching and automated sports analysis for better preservation of traditional folk sports, human pose estimation often needs to be performed on mobile devices with limited computing resources. In this paper, we propose a lightweight human pose estimator termed LiPE. LiPE adopts a lightweight MobileNetV2 backbone for feature extraction and lightweight depthwise separable deconvolution modules for upsampling. Predictions are made at a high resolution with a lightweight prediction head. Compared with the baseline, our model reduces MACs by 93.2 %, and reduces the number of parameters by 93.9 %, while the accuracy drops by only 3.2 %. Based on LiPE, we develop a real- time human pose estimation and evaluation system for automated pose analysis. Experimental results show that our LiPE achieves high computational efficiency and good accuracy for application on mobile devices.
Current human pose estimation models adopt heavy backbones and complex feature enhance- ment modules to pursue higher accuracy. However, they ignore the need for model efficiency in real-world applications. In real-world scenarios such as sports teaching and automated sports analysis for better preservation of traditional folk sports, human pose estimation often needs to be performed on mobile devices with limited computing resources. In this paper, we propose a lightweight human pose estimator termed LiPE. LiPE adopts a lightweight MobileNetV2 backbone for feature extraction and lightweight depthwise separable deconvolution modules for upsampling. Predictions are made at a high resolution with a lightweight prediction head. Compared with the baseline, our model reduces MACs by 93.2 %, and reduces the number of parameters by 93.9 %, while the accuracy drops by only 3.2 %. Based on LiPE, we develop a real- time human pose estimation and evaluation system for automated pose analysis. Experimental results show that our LiPE achieves high computational efficiency and good accuracy for application on mobile devices.
通讯机构：
[Deng, K ] U;Univ South China, Sch Phys Educ, Hengyang 421001, Peoples R China.
关键词：
Folk sports tourism;investment decision;demand uncertainty;differential game
摘要：
More research should shed light on discovering the optimal investment strategy for folk sports tourism destination (FSTD) projects. Therefore, in this paper, we develop a dynamic game model of FSTD considering the dynamic characteristics of FSTD investment, the mode of division of labor and cooperation between public and private operators, and the uncertainty of consumer demand. Public capital is responsible for constructing infrastructures such as venues, and private capital is responsible for services such as catering and accommodation. To promote the development of the FSTD project, the higher-level government subsidizes public investment. Consumer demand for the program is affected by factors such as the size of the two types of capital, the price and quality of services, and demand uncertainty. The study finds that the subsidy leads to an increase in the quantity of public investment and consumption demand, but private sector investment and the prices of both public and private projects are unaffected by the subsidy; the public sector’s net return varies in an inverted U-shape with the rate of subsidy, but the private sector’s net return rises monotonically. Demand disturbances widen the gap in the net returns of operators between the subsidized and unsubsidized scenarios.
摘要：
A dynamic cooperation is poised to redefine the limits of athlete safety and performance optimization in the dynamic field of sports science. A new age in sports analysis is promised by the combination of artificial intelligence (AI) and the internet of things (IoT), one in which data-driven insights not only improve our comprehension of athletic performance but also aid to reduce hazards. This academic work explores the complex interactions between AI and IoT in the context of sports. The IoT and AI integration appear to be a strong mix that has the potential to redefine the standards for athlete safety and performance improvement. This study explores the complex interactions between AI and IoT in the field of sports, emphasizing their combined potential for identifying risk factors in a variety of fields. There is a chance to proactively solve sports-related difficulties by utilizing the data-driven capabilities of IoT and the analytical power of AI, opening the door for better informed tactics and decision-making. Through an exploration of this symbiotic relationship, this paper seeks to underline the transformative potential of these technologies in fostering a safer and more performance-oriented sports environment.
A dynamic cooperation is poised to redefine the limits of athlete safety and performance optimization in the dynamic field of sports science. A new age in sports analysis is promised by the combination of artificial intelligence (AI) and the internet of things (IoT), one in which data-driven insights not only improve our comprehension of athletic performance but also aid to reduce hazards. This academic work explores the complex interactions between AI and IoT in the context of sports. The IoT and AI integration appear to be a strong mix that has the potential to redefine the standards for athlete safety and performance improvement. This study explores the complex interactions between AI and IoT in the field of sports, emphasizing their combined potential for identifying risk factors in a variety of fields. There is a chance to proactively solve sports-related difficulties by utilizing the data-driven capabilities of IoT and the analytical power of AI, opening the door for better informed tactics and decision-making. Through an exploration of this symbiotic relationship, this paper seeks to underline the transformative potential of these technologies in fostering a safer and more performance-oriented sports environment.
期刊：
International Journal of Environmental Research and Public Health,2023年20(4):3140- ISSN：1661-7827
通讯作者：
Xinghong Dai
作者机构：
[Zhiling Chen; Zhigang Tan] School of Physical Education, University of South China, Hengyang 421001, China;School of Physical Education, Hunan University, Changsha 410082, China;Author to whom correspondence should be addressed.;[Xinghong Dai] School of Physical Education, Hunan University, Changsha 410082, China<&wdkj&>Author to whom correspondence should be addressed.
通讯机构：
[Xinghong Dai] S;School of Physical Education, Hunan University, Changsha 410082, China<&wdkj&>Author to whom correspondence should be addressed.
期刊：
MOBILE NETWORKS & APPLICATIONS,2023年28(6):2204-2214 ISSN：1383-469X
通讯作者：
Deng, BY
作者机构：
[Deng, Baiyun; Ren, Lili; Deng, BY] Univ South China, Solux Coll Architecture & Design, Hengyang 421001, Hunan, Peoples R China.;[Chen, Zhiling] Univ South China, Sch Phys Educ, Hengyang 421001, Hunan, Peoples R China.
通讯机构：
[Deng, BY ] U;Univ South China, Solux Coll Architecture & Design, Hengyang 421001, Hunan, Peoples R China.
关键词：
Energy utilization;Internet protocols;Routing protocols;Sensor nodes;Battery cost;Dynamic host configuration protocols;Ecological cultures;Internet of thing;Media communications;New media;New medium communication;Routing-protocol;Tourism industry;Wireless network sensors;Internet of things
