CARSLIDE: T-MAS: : Automatic Tennis Game Analysis

ABSTRACT

Sports video research is a popular topic that has been applied to many prominent sports for a large spectrum of applications. In this project we introduce a technology platform which has been developed for the tennis context, able to capture and evaluate tennis players’ performance during training and official matches.

 

Example of real-time identification of ball and player on a grayscale image

 

Virtual hardware setup for evaluating overlapped fields of view

 

Example of processed action, from a serve to a point

 

Example of processed action with highlighted key images

1
1

PRINCIPAL INVESTIGATOR
ETTORE STELLA

   TYPE AND DATE
   PROVAT CONTRACT
   STARTING DATE – 01/01/2013
   ENDING DATE – 31/12/2015

COST
TOTAL COST €:
TOTAL FUNDING €: 95.000

CNR ISSIA
COST € 
FUNDING €

SUMMARY

Project T-MAS aims to realize a prototype for semi-automatic analysis of situations during a tennis match by means of computer vision techniques. High performance cameras are arranged in a dedicated stereo vision hardware setup that covers a tennis court. System implementation also requires specialized software to acquire, store and process high temporal resolution images in real-time, enabling a software agent to effectively perform 3D reconstruction and scene understanding during matches. Three dimensional positions of active entities such as balls and players are identified, processed as semantically-valid trajectories and then stored on an appropriate database. This technology platform can then exploit pattern recognition and machine learning techniques to help tennis coaches and players improving their tennis performance.

http://www.issia.cnr.it/wp/wp-content/themes/salient-2/css/fonts/svg/5.svg

SEE ALL SENSING, COMMUNICATION AND COMPUTATIONAL INTELLIGENCE PROJECTS

SOURCE OF FOUNDING

PARTNERSHIP