Arnis Performance Evaluation for Beginners and Intermediate: Utilizing SVM Algorithm and Kinect V2 Sensor for Dynamic Stance and Movements

This study investigates the development and evaluation of a Kinect-based system for evaluating the precision of Arnis martial arts techniques. The device employs a calibration method to guarantee accurate measurement of body joint angles during the execution of strikes. Participants perform strikes according to selected difficulty levels (beginner or intermediate) while receiving immediate feedback. To ascertain the efficacy of the system, a statistical analysis will be conducted using Pearson's correlation coefficient to evaluate the level of concurrence between the system's ratings and assessments made by professional trainers. This system offers a promising tool for quantifying and assessing Arnis strikes, potentially providing valuable feedback for beginners and intermediate practitioners. The system's accuracy was verified by comparing it with assessments supplied by professional trainers. The results demonstrated an impressive recognition rate of 89.30% in accurately recognizing the correct form for Arnis martial arts techniques. Future research with a larger participant pool and exploration of additional strike evaluation metrics could further enhance the system's robustness and applicability in Arnis training.