Virtual Tennis Coach using Pose Estimation

Taha Ali, Yashwant Yadav, Sunny Jayant, Jasraj Meena

Abstract: Tennis enjoys a considerable following in India, although it is limited to urban areas but still it is counted among the most popular sports in India. India has produced a number of tennis players, who have achieved international recognition. However, Coaching for this sport is expensive making it unaffordable for many. A method for creating virtual coaches is presented which uses tenser flow posenet algorithm. A database is developed which contains the images of top tennis players. After applying machine learning algorithms on it, the data is able to segregate good posture from the bad ones. The player is provided with an option to compare his game with the game of top ATA players and the end result is shown in visual format to make it more user friendly. The system works on real time data.

Keywords: tenser flow, pose net algorithm, machine learning, ATA players, virtual coach, tennis