New Real-Time Approaches for Video-Genre-Classification Using High-Level Descriptors and a Set of Classifiers
Ronald Glasberg, Sebastian Schmiedeke, Martin Mocigemba, Thomas Sikora · 2008
In this paper we describe in detail the recent publications related to video-genre-classification and present our improved new approaches for classifying video sequences in real-time as 'cartoon', 'commercial', 'music', 'news' or 'sport' by analyzing the content with new high-level audio-visual descriptors and classification methods. Such applications have also been discussed in the context of MPEG-7. The results demonstrate identification rates of more than 90% based on a large representative collection of 100 videos gathered from free digital TV and Internet.