Paper Info
Title
Experiments in the Application of Computer Vision for Ball and Event Identification in Indoor Sports.
Abstract
There is little evidence of thorough experimental research having been performed for the investigation of optimum computer vision algorithms suitable for table tennis umpire support systems and player performance analysis. The only practical way of evaluating and comparing such algorithms is to gather source video data from repeatable experiments in controlled conditions and pass multiple sequences through the same validated algorithms, techniques and processes. This paper presents a digest of software engineering issues, considerations and challenges observed as a direct result of performing over 250 experiments with computer vision and table tennis. Results demonstrate that variations in camera configuration have a direct impact on the success of a selected algorithm and that a comparison of algorithms against the source video data must be made for confirmation. Finally, determination of the location of the ball itself varies between algorithms and a parallel, non-obtrusive technology is required to provide baseline data.
Year
DOI
Venue
2014
10.1007/978-3-319-13153-5_27
COMPUTATIONAL INTELLIGENCE IN INFORMATION SYSTEMS
Keywords
DocType
Volume
object segmentation,detection,ball,sports,computer vision,table tennis,algorithm,comparison
Conference
331
ISSN
Citations 
PageRank 
2194-5357
0
0.34
References 
Authors
0
4
Name
Order
Citations
PageRank
Kevin M. Oldham100.34
Paul Wai Hing Chung231.08
E. A. Edirisinghe39920.09
Benjamin J. Halkon400.68