| dc.description.abstract | Human identification is an important issue in identity authentication which can be applied in many applications, such as security monitoring system, ATM authentication, and personal authentication in businesses transactions. There are many mature image-based human identification techniques that have been developed, such as fingerprints, face, and iris biometric modalities. However, these methods impose severe constraints, such as requiring of a cooperative subject, views from certain aspects, and physical contact or close proximity. To relieve these constraints, human gait identification is a new choice to remedy the problems.
The existing human gait identification methods, including GEI、GHI、GMI…etc, are formed by combining the whole human gait cycle into one image. In this thesis, an effective human gait identification method is presented by separating one cycle into 4 cycles via different combination methods. Experimental results reveal the feasibility and effectiveness of the proposed method in gait identification. We also compare the performance of GEI and FDHI and through experimenting to explain the effects of the four different cycles. The results confirm that our proposed FDHI identification is better than GEI identification.
| en_US |