Badminton is a sport that is very popular in Asia for players of all gender and all ages. The practice of basic badminton skill such as badminton posture practice is essential because the correct posture help to avoid injury and help for improving player skill. In this article, we propose The Model for Practice Badminton Basic Skills. We have created the video posture embedding by using the Triplet-Loss technique and develop the badminton player's motion posture detection by using the One-Shot Learning technique. The motion posture detection consists 8 badminton's postures and 4 other postures instances 1) Forehand Clear 2) Backhand Clear 3) Forehand Drop Shot 4) Backhand Drop Shot 5) Forehand Smash 6) Backhand Smash 7) Forehand Serve 8) Backhand Serve 9) Raise the right hand 10) Raise the left hand 11) Stand/Walk and 12) Run. The performance of the model present the Precision is 0.856, Recall is 0.845, F-Measure is 0.847 and the Accuracy is 0.845
© Copyright 2019 AICCC 2019: Proceedings of the 2019 2nd Artificial Intelligence and Cloud Computing Conference December 2019. Alle Rechte vorbehalten.