Face Recognition Based on Vector Quantization Building Feature Frequency Database

The human faces recognition has generated much research interest nowadays. However, human faces are very similar in structure with minor differences from person to person. Furthermore, lighting condition changes, facial expressions, and hairstyle variations further complicate the face recognition task as one of the difficult problems in pattern analysis. This paper proposes a new face recognition method. Extract the face information from the image and eliminate the influence of the lighting and hair. Use vector quantization (VQ) building the feature frequency database about the face information for recognition. It proved by a lot of experiments that this method has a simple algorithm, high recognition ratio. And for the different brightness, expression and hairstyle it has robustness in a certain extent. The promising results clearly demonstrate the effect of this method. Building a feature database with 30 persons, use the images with different brightness, hairstyle and expression to recognize, and the recognition ratio should achieve 97.6%. It is better than the traditional Fisherfaces and eigenfaces recognition methods.