An Algorithm for Measuring Semantic Similarity of Video Story Unit Research
Wei Wei · Journal of Chengdu University of Information Technology · 2013
Video semantic similarity is applied in many areas of video processing technology.Most video similarity researches were based on low-level features,which caused the semantic gap problem between machine understanding and human thought.In this article,a video story unit semantic model is prompt to try to bridge the low-level features and high-level semantics based on multimodal fusion and multilevel analysis technique with the help of manual label to pre-process the video data.Then an algorithm is come up with it to compute the similarity of two videos.The experiment results of different types of videos indicate that the semantic model can represent the semantic information efficiently,and the algorithm can calculate the similarity of the videos more accurately to reduce the semantic gap.