2016 年 15 巻 6 号 p. 645-649
In order to realize the use of ear biometrics to support criminal investigations, it is essential to study the methods for adjusting the difference in the shooting angles between the security camera and suspect database images. To make up the difference between the shooting angles, previous works have attempted to estimate the feature values of different shooting angles by using multivariate analysis and machine learning techniques. In this study, we propose a method to obtain the feature values using a modified Gabor filter; moreover, we conduct comparative experiments between our proposed method and the conventional method using only using multivariate analysis and machine learning techniques. The experimental results demonstrate that our proposed method achieves an improvement of approximately two percent in terms of equal error rate.