원문정보
초록
영어
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.
목차
1. Introduction
2. Extraction of Multimode Semantic Clues
2.1. Scoring Board (SB)
2.2. Referee Ratio (RR)
2.3. Frame Motion (FM)
3. Detection of Highlights Based on Multimode Semantic Clues and HCRF Model
3.1. Definition of HCRF Model
3.2. Training and Reasoning of HCRF Model
4. Experimental Analysis and Results
4.1. Experiment on Retrieving Multimode Semantic Clues
4.2. Experiment on Detecting Wonderful Events
6. Conclusion
Acknowledgements
References