期刊名称：International Journal of Multimedia and Ubiquitous Engineering
印刷版ISSN：1975-0080
出版年度：2015
卷号：10
期号：10
页码：47-56
DOI：10.14257/ijmue.2015.10.10.05
出版社：SERSC
摘要：A new framework based on multi- modal semantic clues and HCRF (Hidden Conditional Random Field) for soccer wonderful event detection. Through analysis of the structural semantics of the wonderful event videos, define nine kinds of multi- modal semantic clues to accurately describe the included semantic information of the wonderful events. After splitting the video clips into several physical shots, extract the multi- modal semantic clues from the key frame of each shot to get the feature vector of the current shots, and compose the observed sequence of the feature vectors of all shots in the test video clips. Using the above observed sequence as HCRF model input in the case of small-scale training samples, establish wonderful event detection HCRF model effectively. Experiments show that the average recall of this paper reaches 95.3%, the average precision rate reaches 96%, the performance of this paper is obviously superior to the contrast method.
关键词：pixel unmixing; end-member extraction; data field; weighted posterior probability support vector machine
