Quantitative similarity computing for audio effect semantic in video content analysis

As a result of recent progress in networks and hardware, digital video data are increasing significantly. Audio in the video carries abundant semantic message. To understand audio semantic, an audio track semantic similarity computing approach for video semantic content analysis is proposed in this paper. In order to provide a uniform way to make different audio effects to compute quantify similarity, the recognized audio effect will be represented in WordNet, a knowledge bodies that providing a sharing conceptualization. Then, semantic distance in WordNet of two audio effects is used to measures of similarity. The results of experiments comparing to semantic recognition method directly using probability indicate the basic audio semantic analysis method could extraction semantic effectively.