Video retrieval based on a new distance of histograms

Summarizing and understanding video shot based on their contents is an important research topic in multimedia data mining. People tend to search their interested shots from a plenty of videos and many algorithms are used in this field. The algorithm based on histograms is one basic technology of existing methods, however, traditional histograms have the limit on similarity measurement and usually ignore the statistical information of feature. This paper presents an efficient algorithm based on histograms. A novel similarity measurement between histograms is also introduced. The experimental results show that our algorithm performs well and can play an important role in content-based video shot retrieval.