Techniques for designing a classifier for multimedia indexing

Abstract

This paper addresses the issues involved in designing a classifier for multimedia indexing, a representative of domain of tasks involving high dimensionality of feature space and large dissimilarity between features in range and variation, and requiring a strong inference mechanism. We consider decision trees, bayesian network, neural network and support vector approaches The Modified Bayesian Network (MBN), as designed by us offers significant advantages over other approaches. The application of bayesian network has generally been restricted to domains having discrete variable values (such as binary), or to the domain with continuous variable values which approximate to Gaussian distribution. However, MBN can form sound representation of non-Gaussian Multimodal continuous distribution, as is the case with feature space in multimedia indexing. This can be accomplished by intelligent partitioning and data clique association. The structure of MBN and its functionality on real video is also presented in the paper. MBN can perform optimal classification even with partially specified queries given by the user. The strategy automatically gives more weightage to the relevant features amongst hundreds of features present in multimedia indexing system. The inference mechanism is based on iteratively comparing the posterior probability of one class with all other classes. In the comparison of one class with another class, each feature takes different importance measure corresponding to the discerning capacity of the feature. A label is assigned to the multimedia data corresponding to the winning class. The comparison with other classification tools shows that MBN classification performance is consistently better than that of the other tools.