The design of a recognition system requires careful attention to pattern representation and classifier design. Some statistical approaches choose these features, in a higher dimensional initial space, which allow sample vectors belonging to different categories to occupy compact and disjoint cluster regions in a lower dimensional subspace. The effectiveness of the subspace is determined by how well samples from different classes can be separated. This paper describes a feature selection process for a pose invariant 3D facial expression recognition method providing a lower dimensional subspace representation, which is optimized to improve the classification accuracy, retrieved from geometrical localization of facial feature points to classify universal facial expressions. Probabilistic neural network architecture is employed as a classifier to recognize the facial expressions from the feature vectors obtained from 3D facial feature locations. Facial expressions such as Neutral, Anger, Disgust, Fear, Happiness, Sadness, and Surprise are successfully recognized with an average recognition rate of 93.72%.
展开▼
