Achieving semantic coupling in the domain of high-dimensional video indexing application

Abstract

A typical user of multimedia indexing system has to deal with low-level details like assignment of weights to different features and selection of threshold. To employ the system effectively and efficiently, one has to be versatile in similarity measures used, and other implementation issues. In order to overcome these shortcomings, researchers in the field of content-based retrieval are making efforts to develop semantic features. However, most of the work done in semantic features indexing build a priori model of a video's structure that is based on domain knowledge; and they are focused on narrow domains like extracting interesting shots from 'soccer', finding 'Tennis' shots etc. In this paper, an adequately domain-independent approach is presented where local features can characterize multimedia data using Neural Networks (ANN) and Support Vector Machines (SVM). In our previous work, we have shown that classification in content-based retrieval requires non-linear mapping of feature space. This can normally be accomplished by ANN and SVM. However, they inherently lack the capability to deal with meaningful feature evaluation and large dimensional feature space in the sense that they are inaccurate and slow. These defects can be overcome by employing meaningful feature selection on the basis of discriminative capacity of a feature. The experiments on database consisting of real video sequences show that the speed and accuracy of SVM can be improved substantially using this technique, while execution time can be substantially reduced for ANN. The comparison also shows that improved SVM turns out to be a better choice than ANN. Finally, it is shown that generalization in learning is not affected by reducing the dimension of the feature space by our method.