[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-458130-105":3,"detail-sidebar-cat-0-en-105":80,"doc-detail-458130-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":73,"head_meta":75,"extra_data":77,"updated_unix":79},105,"en","development-of-user-customized-online-teaching-technology-based-on-gpt-enhanced-by-generative-ai-based-motion-recognition","Development of user-customized online teaching technology based on GPT - enhanced by generative AI-based motion recognition","","A proposal addresses weaknesses in existing vision-based Human Pose Estimation for joint tracking and motion recognition in crowded scenes and under occlusion or special clothing. A Numerical Discriminator Generative Adversarial Network (ND_GAN) improves joint estimation through three integrated modules. Experiments use yoga videos from Hanchoom, comparing MediaPipe vs ND_GAN on original and legmasked data, achieving 94.2% PCK@0.5, 93.1% PCK@0.2, and 92.8% F1. ND_GAN remains stable under occluded conditions and surpasses DWPose and ViTPose by over 15%. A GPT-based teaching system provides real-time feedback on movement errors.",{"@graph":14,"@context":72},[15,34,55],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/technology/","Technology",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/development-of-user-customized-online-teaching-technology-based-on-gpt-enhanced-by-generative-ai-based-motion-recognition/458130/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/development-of-user-customized-online-teaching-technology-based-on-gpt-enhanced-by-generative-ai-based-motion-recognition/458130.png","ImageObject",300,407,{"name":42,"@type":43},"Miles","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-16","2026-09-30",true,{"@type":52,"interactionType":53,"userInteractionCount":33},"InteractionCounter",{"@type":54},"ViewAction",{"@type":56,"mainEntity":57},"FAQPage",[58,64,68],{"name":59,"@type":60,"acceptedAnswer":61},"What problem does the paper target in current motion recognition and pose estimation?","Question",{"text":62,"@type":63},"It targets inaccurate joint tracking and inconsistent motion recognition in crowded spaces and when users wear special attire, especially under occlusion.","Answer",{"name":65,"@type":60,"acceptedAnswer":66},"How does ND_GAN improve Human Pose Estimation performance?",{"text":67,"@type":63},"ND_GAN uses three integrated modules based on a Numerical Discriminator GAN to enable more precise joint estimation in complex environments.",{"name":69,"@type":60,"acceptedAnswer":70},"What are the reported quantitative results compared with MediaPipe and other state-of-the-art models?",{"text":71,"@type":63},"On the evaluated yoga videos, ND_GAN reports 94.2% PCK@0.5, 93.1% PCK@0.2, and 92.8% F1-score, improving by over 30% versus MediaPipe, and exceeding DWPose and ViTPose by more than 15% in accuracy.","https://schema.org",{"og:url":32,"og:type":74,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":76,"canonical":32},"index,follow",{"doc_id":78,"site_id":7},458130,1791513358,{"code":4,"msg":81,"data":82},"success",[83,87,91,95,100,104,109,114,119,122,126],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":84,"show_sort_weight":85,"slug":86},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":88,"show_sort_weight":89,"slug":90},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":92,"show_sort_weight":93,"slug":94},"Exam",70,"exam",{"id":96,"doc_module":4,"doc_module_name":25,"category_name":97,"show_sort_weight":98,"slug":99},5,"Comic",60,"comic",{"id":101,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":102,"slug":103},6,50,"technology",{"id":105,"doc_module":4,"doc_module_name":25,"category_name":106,"show_sort_weight":107,"slug":108},7,"Healthcare",40,"healthcare",{"id":110,"doc_module":4,"doc_module_name":25,"category_name":111,"show_sort_weight":112,"slug":113},8,"Research & Report",30,"research-report",{"id":115,"doc_module":4,"doc_module_name":25,"category_name":116,"show_sort_weight":117,"slug":118},9,"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":25,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":25,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":96,"slug":129},19,"General","general",{"code":4,"msg":81,"data":131},{"doc_id":78,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":101,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":33,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":139,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":112},13056703019404,"https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0","[www. nature.com/scientificreports](www. nature.com/scientificreports)  \nOPEN  \nDevelopment of user-customized online teaching technology based on GPT, enhanced by generative AIbased motion recognition  \nKyoung-Geun Cho1, Zahra-Batool Jaffrey1, Hun-Hee Cho1, Jun-Woo Lee1, Ye-Jin Lee1, Seon Uck Paek1,4, Seo-Young Won2, Zolzaya Dashdorj3, Erdenebaatar Altangerel3 & Tae-Koo Kang1􀀍  \nThis paper addresses current issues in existing joint tracking and motion recognition algorithms for Human Pose Estimation. It proposes a solution using the Numerical Discriminator Generative Adversarial Network (ND_GAN) to improve the performance of vision-based motion recognition technology. Existing algorithms face challenges in accurately tracking joints in crowded spaces or with users wearing special attire, resulting in reduced accuracy and inconsistent results. The proposed ND_GAN consists of three integrated modules, enabling more precise joint estimation even in complex environments. Experiments were conducted using yoga videos from Hanchoom, a home training platform by TDI, comparing the performance of MediaPipe and ND_GAN on both original and legmasked videos. Quantitative evaluation showed that ND_GAN achieved 94.2% PCK@0.5, 93.1%  \nPCK@0.2, and an F1-score of 92.8%, marking an improvement of over 30% compared to the existing MediaPipe model. Notably, ND_GAN consistently estimated joint coordinates even in occluded video conditions, demonstrating significant performance gains in scenarios where traditional models failed. Furthermore, ND_GAN outperformed cutting-edge state-of-the-art models such as DWPose and ViTPose by more than 15% in accuracy, confirming its robustness in real-world environments with various occlusion conditions. Additionally, a teaching system using GPT is introduced to provide real-time feedback on movement errors, supporting users in effectively learning the motions.  \nThis integrated approach marks a significant advancement in addressing Human Pose Estimation complexities, enhancing overall vision-based motion recognition efficacy.  \nKeywords Motion recognition, Generative AI, GPT, Teaching system  \nResearch background  \nComputer vision aims to implement human visual perception abilities in computers. Since the outbreak of COVID-19 in 2019, the emphasis on non-face-to-face interactions has increased, shedding more light on the development of computer vision and artificial intelligence technologies.  \nFigure 1 depicts the change in keyword searches using the Google search engine. As shown in Fig. 1, the keywords ‘non-contact’ and ‘contactless’ started gaining attention due to concerns about virus transmission through face-to-face interactions. With the onset of non-face-to-face environments, the consumption of inperson services provided by individuals such as healthcare trainers or choreographers significantly decreased after the emergence of COVID-19, leading to an increase in the proportion of non-face-to-face service consumption1.  \nIn response to this shift, society has actively developed computer vision and artificial intelligence to enable individuals to use services that traditionally required face-to-face interactions in front of mechanical elements like computers. In the context of health care services, trainers aim to observe and assess the precise movements and postures of individuals receiving health care services in face-to-face settings. They provide feedback on movement and posture issues, guiding individuals to use the services correctly. The challenge is providing such  \n1Department of Human Intelligence Robot Engineering, Sangmyung University, Cheonan, Republic of Korea. 2TDI Co., Seoul, Republic of Korea. 3Computer Science Department, Mongolian University of Science and Technology, Ulaanbaatar, Mongolia. 4 Department of Software, Sangmyung University, Cheonan, Republic of Korea. 􀀍 email: [tkkang@smu.ac.kr](tkkang@smu.ac.kr)  \n[www. nature.com/scientificreports/](www. nature.com/scientificreports/)  ","cbCaieUPekIV6gOB","https://ap.wps.com/l/cbCaieUPekIV6gOB","pdf",2310108,12,"English","# Research background\n## Limitations of current motion recognition model algorithms","[{\"question\":\"What problem does the paper target in current motion recognition and pose estimation?\",\"answer\":\"It targets inaccurate joint tracking and inconsistent motion recognition in crowded spaces and when users wear special attire, especially under occlusion.\"},{\"question\":\"How does ND_GAN improve Human Pose Estimation performance?\",\"answer\":\"ND_GAN uses three integrated modules based on a Numerical Discriminator GAN to enable more precise joint estimation in complex environments.\"},{\"question\":\"What are the reported quantitative results compared with MediaPipe and other state-of-the-art models?\",\"answer\":\"On the evaluated yoga videos, ND_GAN reports 94.2% PCK@0.5, 93.1% PCK@0.2, and 92.8% F1-score, improving by over 30% versus MediaPipe, and exceeding DWPose and ViTPose by more than 15% in accuracy.\"}]","Development of user-customized online teaching technology based on GPT - enhanced by generative AI-based motion recognition | PDF",1790751733]