Penjejakan (Tracking) Pemain Bola Menggunakan Deep Learning Pada Rekaman Pertandingan Sepak Bola

Narayana, Komang Alit (2024) Penjejakan (Tracking) Pemain Bola Menggunakan Deep Learning Pada Rekaman Pertandingan Sepak Bola. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5024201066-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (3MB) | Request a copy

Abstract

Sepak bola adalah olahraga yang kompleks dengan pergerakan pemain yang cepat dan beragam. Analisis pergerakan pemain dalam pertandingan dapat memberikan wawasan penting untuk meningkatkan strategi tim. Tugas akhir ini bertujuan untuk mengembangkan sistem penjejakan pemain bola dalam rekaman pertandingan sepak bola dengan memanfaatkan pendekatan deep learning. Metodologi tugas akhir ini mencakup pengumpulan rekaman pertandingan sepak bola, pelabelan data untuk mengidentifikasi posisi pemain, dan pengembangan model deep learning yang mampu mengenali dan melacak pemain. Hasil evaluasi kinerja model menggunakan metrik seperti precision, recall, MAP-50 (Mean Average Precision at IoU threshold 0.5), dan MAP 50-95 (Mean Average Precision di rentang IoU threshold 0.5 hingga 0.95) yang menunjukkan nilai akurasi dalam penjejakan pemain. hasil penjejakan pemain yang dilakukan diharapkan nantinya dapat membantu dalam menganalisa suatu pertandingan sepak bola, serta dapat memberikan manfaat besar bagi pelatih dalam menyusun sebuah strategi kedepannya. Tugas akhir ini berkontribusi pada pengembangan teknologi dalam dunia olahraga, membuka potensi untuk meningkatkan kualitas dan hasil pertandingan sepak bola melalui pemahaman yang lebih mendalam tentang pergerakan pemain melalui deep learning
===================================================================================================================================
Football is a complex sport with fast and varied player movements. Analysis of player movements in a match can provide important insights to improve team strategy. This final project aims to develop a soccer player tracking system in soccer match recordings by utilizing a deep learning approach. The methodology for this final project includes collecting soccer match recordings, labeling data to identify player positions, and developing a deep learning model that is able to recognize and track players. The results of the model performance evaluation use metrics such as precision, recall, MAP-50 (Mean Average Precision at IoU threshold 0.5), and MAP 50-95 (Mean Average Precision in the IoU threshold range 0.5 to 0.95) which shows the accuracy value in player tracking. It is hoped that the results of the player tracking carried out will later be able to help in analyzing a football match, and can provide great benefits for coaches in developing a strategy in the future. This final project contributes to the development of technology in the world of sport, unlocking the potential to improve the quality and results of football matches through a deeper understanding of player movement through deep learning

Item Type:	Thesis (Other)
Uncontrolled Keywords:	: Analisa, Deep Learning, Metrik, Sepak Bola, Teknologi; Analysis, Deep Learning, Metrics, Soccer, Technology
Subjects:	T Technology > T Technology (General) > T385 Visualization--Technique T Technology > T Technology (General) > T57.62 Simulation T Technology > T Technology (General) > T58.8 Productivity. Efficiency
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	Komang Alit Narayana
Date Deposited:	25 Jul 2024 03:20
Last Modified:	25 Jul 2024 03:20
URI:	http://repository.its.ac.id/id/eprint/108778

Actions (login required)

View Item