원문정보
보안공학연구지원센터(IJMUE)
International Journal of Multimedia and Ubiquitous Engineering
Vol.11 No.1
2016.01
pp.33-42
피인용수 : 0건 (자료제공 : 네이버학술정보)
초록
영어
Football is one of the most popular sports in the world. It is deeply loved by massive ball fans. Generally one complete football game lasts around two hours, but only a very little part of it has wonderful views which attract audiences. Although in sport matches, episodes which people care about are of high subjectivity, for football games, they share some in common. . So how to fetch out concerned events from tremendous football videos has become a very important and typical problem. Here we propose a new algorithm for extracting effectively key events from football videos, which is named key event extraction method based on attributes of shots. The method proves higher recall and precision rate.
목차
Abstract
1. Introduction
2. Goal Event Extraction Method Based on Attributes of Shots
2.1 Creation of Rules
2.2. Determination of Attributes of Relevant Shots
2.3. Implementation of the Algorithm
3. Experimental Analysis and Results
3.1. Detection Results of Shots in Football Videos
3.2. Determination of Potential Goal Views
3.3. Detection Results of Slow-Lens Replay
3.4 Extraction Results of Goal Events
4. Conclusion
References
1. Introduction
2. Goal Event Extraction Method Based on Attributes of Shots
2.1 Creation of Rules
2.2. Determination of Attributes of Relevant Shots
2.3. Implementation of the Algorithm
3. Experimental Analysis and Results
3.1. Detection Results of Shots in Football Videos
3.2. Determination of Potential Goal Views
3.3. Detection Results of Slow-Lens Replay
3.4 Extraction Results of Goal Events
4. Conclusion
References
저자정보
참고문헌
자료제공 : 네이버학술정보