News Story Detection Method Based on Multi-feature Fusion
Yang Liu · Journal of Chinese Computer Systems · 2008
It is important to analyze the structures of news videos and to understand their semantic content,news story boundary detection is one effective way of solving the above problems.By the preliminary analysis of news structures,a news story detection method based on multi-feature fusion is proposed.The method avoids the deficiency of detecting story boundaries by a single feature,but fuses multi-features such as shots,anchorpersons,titles and silences,etc.Experiments prove the effectiveness of the method,which can get the average detection precision of 96%,and can accomplish the task of detecting story boundaries.