Maximum-likelihood dimensionality reduction in gaussian mixture models with an application to object classification

Accurate classification of objects of interest for video surveillance is difficult due to occlusions, deformations and variable views/illumination. The adopted feature sets tend to overcome these issues by including many and complementary features; however, their large dimensionality poses an intrinsic challenge to the classification task. In this paper, we present a novel technique providing maximum-likelihood dimensionality reduction in Gaussian mixture models for classification. The technique, called hereafter mixture of maximum-likelihood normalized projections (mixture of ML-NP), was used in this work to classify a 44-dimensional data set into 4 classes (bag, trolley, single person, group of people). The accuracy achieved on an independent test set is 98% vs. 80% of the runner-up (MultiBoost/AdaBoost).