Мазурець, Олександр ВікторовичСобко, Олена ВіталіївнаБогданова, Ангеліна ГригорівнаMazurets, O.Sobko, O.Bohdanova, A.2023-09-132023-09-132023Bohdanova A., Mazurets O., Sobko O. Gesture recognition using a neural network in real time. Black Sea Science 2023: Proceedings of the International Competition of Student Scientific Works. Odesa National University of Technology. Odesa, ONUT, 2023. Pp. 556-566https://elar.khmnu.edu.ua/handle/123456789/14417The aim of this work is to develop and implement a method for gesture recognition. The main functions of the application are capturing images from a web camera, memorizing and recognizing user gestures, and training the created model. Javascript programming language and TensorFlow.js, Jest, Bootstrap, Three.js libraries were used for development. The developed neural network is designed to recognize user gestures, namely recognizing digits formed by hand gestures. The practical use of the developed neural network is determined by capturing the current user's gestures in real time from the video stream, searching for the corresponding gesture of the digit and facilitating user interaction when working with the computer, which is especially important for people with limited abilities.The aim of this work is to develop and implement a method for gesture recognition. The main functions of the application are capturing images from a web camera, memorizing and recognizing user gestures, and training the created model. Javascript programming language and TensorFlow.js, Jest, Bootstrap, Three.js libraries were used for development. The developed neural network is designed to recognize user gestures, namely recognizing digits formed by hand gestures. The practical use of the developed neural network is determined by capturing the current user's gestures in real time from the video stream, searching for the corresponding gesture of the digit and facilitating user interaction when working with the computer, which is especially important for people with limited abilities.enNeural networkgesture recognitiondigit recognitionreal-time recognition