Object detection in basketball game videos
Αναγνώριση αντικειμένων σε βίντεο αγώνων μπάσκετ
View/
Keywords
Αναγνώριση αντικειμένωνAbstract
Object detection in sports analytics has gained significant attention due to its potential to enhance
game analysis, player performance tracking, and automated officiating. This thesis focuses
on developing an advanced object detection system specifically designed for basketball
game videos. The proposed approach utilizes deep learning techniques, particularly the YOLOv8
model, to detect and track key elements such as players, the ball, and the hoop. Additionally,
homography transformation is applied to map player positions onto a 2D court representation, enabling
more effective movement tracking and spatial analysis. The dataset used is sourced from
real-game footage and preprocessed using court segmentation techniques to improve detection
accuracy. This research contributes to the growing field of sports analytics by providing an automated,
real-time object detection system that can assist coaches, analysts, and broadcasters in
gaining deeper insights into game dynamics.
Postgraduate Studies Programme
ΠληροφορικήDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
69Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα