Soccer video summarization using enhanced logo detection

  • Authors:

  • Mohamed Y. Eldib;Bassam S. Abou Zaid;Hossam M. Zawbaa;Mohamed El-Zahar;Motaz El-Saban

  • Affiliations:

  • Information Technology department, Faculty of Computers and Information, Cairo University, Cairo, Egypt;Information Technology department, Faculty of Computers and Information, Cairo University, Cairo, Egypt;Information Technology department, Faculty of Computers and Information, Cairo University, Cairo, Egypt;Information Technology department, Faculty of Computers and Information, Cairo University, Cairo, Egypt;Information Technology department, Faculty of Computers and Information, Cairo University, Cairo, Egypt

  • Venue:
  • ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

We propose an automatic soccer video summarization engine which relies on an improved algorithm for the detection of replay shots which delineate interesting events. Video shots are first detected using dominant color and histogram intersection methods. Replay shots are detected using an improved technique, then processed through a set of mid-level descriptors (goal-mouth and score board with other cinematic features) and finally fed into a rule-based classifier. The proposed summarization has been experimentally tested on 6 hours from 10 videos in total. The proposed logo-based replay detection technique achieves 100% recall with 96.3% precision. Interesting events such as goals are detected with 100% recall and 100% precision, attacks with 91.7% recall and 86.7% precision) and other events such as (fouls, free kicks etc) with 90.8% recall and 95% precision.