EAR RECOGNITION USING BI-ORTHOGONAL AND GABOR WAVELET-BASED REGION COVARIANCE MATRICES
Ali Pour Yazdanpanah, Karim Faez · Applied Artificial Intelligence · 2010
Here we propose bi-orthogonal and Gabor wavelet-based region covariance matrices as a novel method, which is robust to changes in illumination and pose variations, for ear recognition. In this method we construct a region covariance matrix by using bi-orthogonal and Gabor wavelet features and illumination intensity and pixel location and use it as an efficient and robust ear descriptor. We performed our experimental studies comparing various ear recognition methods, including our proposed method, with the PCA + RBFN method, the ICA + RBFN method, the Hmax + SVM method, the LSBP method, conventional RCM-based method, and the GRCM method. Superiority of our new method has been successfully tested on ear recognition using 488 images corresponding to 137 subjects from two databases 1 and 2 in USTB database. Our proposed method achieves the average accuracy of 96.6% and 93.5%, respectively, on the databases 1 and 2 for ear recognition.