Ear recognition based on weighted wavelet transform and DCT

Feature extraction is the key to improve the ear recognition. Firstly, we conduct two dimensional discrete wavelet transform on the human ear images, and then we implement block discrete cosine transform on the low frequency components of wavelet transform and weighted high-frequency components in order to extract DCT coefficients of the image and construct the feature vectors. Finally, we use the nearest neighbor classifier combined with weighted distance for classification and recognition.. The experimental results show that the new method has a higher recognition rate compared with the method in which only the low frequency components of wavelet transform are used.