AI pakalpojums fitnesa centru uzņēmumiem un apmeklētājiem

Abstract

Daudzi cilvēki katru dienu apmeklē sporta zāles, taču daudzi no cilvēkiem vingrošanās laikā nodara kaitējumu savam ķermenim un to neapzinās. Šajā rakstā ir piedāvāts risinājums, kā palīdzēt cilvēkiem izvairīties no bezmērķīgas sevis traumēšanas, veicot fiziskus vingrinājumus sporta zālēs vai jebkur citur. Risinājums ir mākslīgā intelekta iekārta, kas analizē attēlus un video, kur persona veic fizisko vingrinājumu. Ja vingrinājums tiek veikts nepareizi, iekārta nosūta atgriezenisko saiti ar ieteikumiem, kur vingrinājumā ir nepieciešami uzlabojumi. Keywords Artificial Intelligence, sports exercise, feedback mechanism, gym additional

Growing health awareness has led to a continuous rise in gym memberships in Europe, with around 12% of citizens attending in 2019, resulting in an increase in the number of gyms. However, people are getting injured during gym workouts due to a lack of proper education on exercise performance. It could be solved by hiring private coaches for every session, which can be costly for both gym attendees and gym owners. Digitalization offers a solution by providing AI-based systems for gym clients to reduce expenses and for gym owners to increase revenue through digital fitness services, the demand for which has increased since the COVID-19 pandemic. The author's solution is to create a machine that can film and analyze gym clients exercise performance, providing feedback on how to improve and highlighting mistakes through email, app notifications, or a big screen in front of gym clients. This requires machine learning principles and computer image analysis mechanisms. This would help people reduce traumas and expenses. The method used was creation of app using python and python libraries with needed functionality such as body part recognition. After developing of application is was tested on video files where people were performing plank exercise. While doing exercise people were making 4 popular mistakes that statistically many people are doing. This was made on purpose to test if the app is working. Afterwards, results of app were compared with authors manual observations. During study, it showed that AI has the potential to assist on this kind of questions, like providing feedback. However, it is not fully efficient. The technology used in this study showed that a picture recognition system is a crucial part for the feedback mechanism. If there is no precise picture, then there is incorrect feedback, or not precise enough feedback. However, with bigger research and other tools usage, this kind of machine with the big probability is achievable for creation.