3D head model classification using KCDA

In this paper, the 3D head model classification problem is addressed by use of a newly developed subspace analysis method: kernel clustering-based discriminant analysis or KCDA as an abbreviation. This method works by first mapping the original data into another high-dimensional space, and then performing clustering-based discriminant analysis in the feature space. The main idea of clustering-based discriminant analysis is to overcome the Gaussian assumption limitation of the traditional linear discriminant analysis by using a new criterion that takes into account the multiple cluster structure possibly embedded within some classes. As a result, Kernel CDA tries to get through the limitations of both Gaussian assumption and linearity facing the traditional linear discriminant analysis simultaneously. A novel application of this method in 3D head model classification is presented in this paper. A group of tests of our method on 3D head model dataset have been carried out, reporting very promising experimental results.