Free-Throw Prediction in Basketball Sport Using Object Detection and Computer Vision

Gowda, MS and Shindhe, SD and Omkar, SN (2024) Free-Throw Prediction in Basketball Sport Using Object Detection and Computer Vision. In: 8th International Conference on Computer Vision and Image Processing, CVIP 2023, 3 November 2023through 5 November 2023, Jammu, pp. 515-528.

PDF Com_ Com_ Inf_Sci_2024.pdf - Published Version Restricted to Registered users only Download (113MB) | Request a copy

Abstract

The most important things that must be taken care of in a basketball game include teamwork, perfect team coordination, and execution of the skills of the players at the right point of time, without which it will be difficult to win a game. So, a player must realize these things while he is playing, but in order to do that, he must analyze his posture and the way he plays the game. In order to help the players, a system is developed using computer vision techniques which provides better insights into improving games using deep learning algorithms such that the insights can be used in improving the player performance and strategies in individual or team games. This paper provides a deeper knowledge of the kinematic and physiological markers that might better capture athletic performance by looking at the present state-of-the-art AI approaches by analyzing how AI methods and techniques are applied in basketball play. This work mainly concentrates on the Basketball sport, in which different features such as the release angle of the basketball, and the shot predictions are analyzed and tested in real-time. The object detection is performed using the YOLOv5 algorithm and we have obtained a mean average precision of 96.8. Further, the release angle is calculated using the combination of pose landmarks and object detection and has resulted in the optimal angle for a perfect free throw, which is to be within the range of 45 to 60°. Based on the combination of object detection, release angle, and the polynomial regression, shot prediction is performed and has resulted in accurate results for real-time experimental analysis. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

Item Type:	Conference Paper
Publication:	Communications in Computer and Information Science
Publisher:	Springer Science and Business Media Deutschland GmbH
Additional Information:	The copyright for this article belongs to publisher.
Keywords:	Computer games; Computer vision; Deep learning; Forecasting; Human resource management; Object recognition; Sports, (you only look once) YOLOv5; Basketball analyse; Basketball games; Free throws; Objects detection; Polynomial regression; Pose-estimation; Real- time; Release angles; Team coordination, Object detection
Department/Centre:	Division of Mechanical Sciences > Aerospace Engineering(Formerly Aeronautical Engineering)
Date Deposited:	02 Sep 2024 06:50
Last Modified:	02 Sep 2024 06:50
URI:	http://eprints.iisc.ac.in/id/eprint/86023

Actions (login required)

View Item