Face Recognition Based on Grain-Shape Features

摘要

Traditional face recognition method was mainly dependent on single vision features such as color, texture, shape and so on. So the recognition result was not satisfactory.To cure this problem, propose a new face recognition method to get the texture features and shape features of face image based on wavelet transform. The corresponding shape and texture features are then processed by linear discriminant analysis. The PIE face database was used to test the proposed method. The experiment result shows that the proposed method has better recognizing effect and is not sensitive to the pose and expression of human faces. Experimental result also shows that the method is superior to the PCA and DCT method.