Combination recognition of face and ear based on two-dimensional fisher linear discriminant

Due to the impact of the factors such as age, cosmetics, beard and facial expressions, the performance of face recognition drops, while ear has rich and stable physiological features. A biometrics method combining face and ear is presented. Two-Dimensional Fisher Linear Discriminant (2DFLD) is applied respectively to recognition combining of facial and otic images and recognition combining of facial and otic features. On USTB ear database and ORL face database, the experimental results have shown that the recognition rate of combining facial and otic images reaches 97.5%, and is increased respectively 12.5% and 5% higher than that of face or ear; the recognition rate of combining facial and otic features is 95%, and is increased 10% and 2.5% higher than that of face or ear. The recognition of combination based on the 2DFLD method has a good recognition performance compared with PCA or 2DPCA. This indicates that the combined recognition of multi-biological features is an effective method of biometrics.