摘要
This paper presents two video copy detection systems built for the TRECVID 2010 content-based copy detection task. Three runs were submitted using video-only content. Two systems differ in terms of the feature design as well as the matching scheme. In this paper we overview the underlying methodologies and discuss the various design choices for developing a practical video copy detection system.
原文 | 英語 |
---|---|
出版狀態 | 已發佈 - 2010 |
事件 | TREC Video Retrieval Evaluation, TRECVID 2010 - Gaithersburg, MD, 美国 持續時間: 2010 11月 15 → 2010 11月 17 |
其他
其他 | TREC Video Retrieval Evaluation, TRECVID 2010 |
---|---|
國家/地區 | 美国 |
城市 | Gaithersburg, MD |
期間 | 2010/11/15 → 2010/11/17 |
ASJC Scopus subject areas
- 電腦繪圖與電腦輔助設計
- 電腦視覺和模式識別
- 人機介面
- 軟體