Semantic concept detection for video based on extreme learning machine

Semantic concept detection is an important step in concept-based semantic video retrieval, which can be regarded as an intermediate descriptor to bridge the semantic gap. Most existing concept detection methods utilize Support Vector Machines (SVM) as concept classifier. However, there are several drawbacks of using SVM, such as the high computational cost and large number of parameters to be optimized. In this paper we propose an Extreme Learning Machine (ELM) based Multi-modality Classifier Combination Framework (MCCF) to improve the accuracy of semantic concept detection. In this framework: (i) three ELM classifiers are trained by exploring three kinds of visual features respectively, (ii) a probability-based fusion method is then proposed to combine the prediction results of each ELM classifier, (iii) we integrate the prediction results of ELM classifier with the information of contextual correlation among concepts to further improve the accuracy of semantic concept detection. Experiments on the widely used TRECVID datasets demonstrate that our approach can effectively improve the accuracy of semantic concept detection and achieve performance at extremely high speed.