Identifikasi Teknik Pukulan Yang Digunakan Pada Permainan Bola Voli Menggunakan Yolo V8 Pose Estimation Dan Lstm

Abstract

Volleyball involves a variety of dynamic hitting techniques, requiring intelligent approaches for accurate recognition and analysis. This study aims to develop a system for identifying volleyball hitting techniques using a computer vision-based method. The system utilizes YOLOv8 for body pose estimation and Long Short-Term Memory (LSTM) for temporal classification of movement sequences. The research methodology includes data collection, model training, system implementation, and performance evaluation. The results show that the developed system can identify hitting techniques with an accuracy of 94%, correctly classifying 235 out of 250 test samples. Performance evaluation using precision, recall, and f1-score metrics demonstrates consistent and reliable results across all technique classes. The implementation of LSTM proves effective in capturing sequential body movement patterns, enabling more accurate and context-aware classification. This study confirms that the combination of YOLOv8 and LSTM is a highly effective approach for analyzing sports techniques, with strong potential for further development in AI-based sports learning and training applications.