dc.contributor | BEȘLIU, Corina | |
dc.contributor.author | PETRANIS, Ana | |
dc.contributor.author | VINOGRADSCHII, Mihail | |
dc.date.accessioned | 2024-10-23T08:50:40Z | |
dc.date.available | 2024-10-23T08:50:40Z | |
dc.date.issued | 2024 | |
dc.identifier.citation | PETRANIS, Ana and Mihail VINOGRADSCHII. Enhancing sports performance analysis: an ai approach for basketball and volleyball. In: Conferinţa tehnico-ştiinţifică a studenţilor, masteranzilor şi doctoranzilor = Technical Scientific Conference of Undergraduate, Master and PhD Students, Universitatea Tehnică a Moldovei, 27-29 martie 2024. Chișinău, 2024, vol. 2, pp. 911-916. ISBN 978-9975-64-458-7, ISBN 978 9975-64-460-0 (Vol.2). | en_US |
dc.identifier.isbn | 978-9975-64-458-7 | |
dc.identifier.isbn | 978 9975-64-460-0 | |
dc.identifier.uri | http://repository.utm.md/handle/5014/28262 | |
dc.description.abstract | Artificial intelligence is increasingly being used in all areas of human activity, from the familiar text editor to the cutting-edge satellite that has just entered Earth's orbit. Modern team sports also need to implement AI to analyze the results of matches, in order to identify the strengths and weaknesses of each player, as well as to develop a right strategy for subsequent games against a specific opponent. Our study investigates the process of data collection and processing for sports analytics using basketball and volleyball games as examples. Data for analysis was sourced from the official FIBA YouTube channel "FIBA - The Basketball Channel" and the Baller TV replay library, which archives matches from various youth sports. The obtained videos were segmented into individual frames. A portion of these frames were manually labeled to create training and validation datasets. The remaining frames formed the unlabeled test dataset, crucial for evaluating the accuracy of the YOLOv8 model chosen as the foundation for this study. Our focus was on identifying players through jersey number recognition, detecting the ball and its location on the court, classifying game situations, and processing the score and timer using OCR technology. The fine-tuned YOLOv8 achieved an accuracy of 93% based on the mAP50-95 metric, which evaluates the overlap between predicted and actual bounding boxes. | en_US |
dc.language.iso | en | en_US |
dc.publisher | Universitatea Tehnică a Moldovei | en_US |
dc.relation.ispartofseries | Conferinţa tehnico-ştiinţifică a studenţilor, masteranzilor şi doctoranzilor = Technical Scientific Conference of Undergraduate, Master and PhD Students: Chişinău, 27-29 martie 2024. Vol. 2; | |
dc.rights | Attribution-NonCommercial-NoDerivs 3.0 United States | * |
dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/3.0/us/ | * |
dc.subject | artificial intelligence | en_US |
dc.subject | sports performance | en_US |
dc.subject | machine learning | en_US |
dc.title | Enhancing sports performance analysis: an ai approach for basketball and volleyball | en_US |
dc.type | Article | en_US |
The following license files are associated with this item: