Feature Extraction of Human Facail Expressions Using Haar Wavelet and Karhunen-Loève Transforms

Abstract

One of the challenging and active research topics in the recent years is Facial Expression. This paper presents the method to extract the features from the facial expressions from still images. Feature extraction is very important for classification and recognition process. This paper involve three stages which contain capture the images, pre-processing and feature extractions. This method is very efficient in feature extraction by applying haar wavelet and KL-T. Cohn-kanade used six expressions of anger, sadness fear, happiness, disgust and surprise. Features that have been extracted from the image of facial expressions were used as inputs to the neural network.