Automatic summarization generation of sports video content has been object of\ngreat interest for many years. Although semantic descriptions techniques have been\nproposed, many of the approaches still rely on low-level video descriptors that render\nquite limited results due to the complexity of the problem and to the low capability\nof the descriptors to represent semantic content. In this paper, a new approach for\nautomatic highlights summarization generation of soccer videos using audio-visual\ndescriptors is presented. The approach is based on the segmentation of the video\nsequence into shots that will be further analyzed to determine its relevance and interest.\nOf special interest in the approach is the use of the audio information that provides\nadditional robustness to the overall performance of the summarization system. For\nevery video shot a set of low and mid level audio-visual descriptors are computed and\nlately adequately combined in order to obtain different relevance measures based on\nempirical knowledge rules. The final summary is generated by selecting those shots\nwith highest interest according to the specifications of the user and the results of\nrelevance measures. A variety of results are presented with real soccer video sequences\nthat prove the validity of the approach.
Loading....