Author
Abstract
The Biometrics recognition systems act as an efficient method with broad
applications in the area of: security access control, personal identification to humancomputer
communication. From other hand, some biometrics have only little variation
over the population, have large intra-variability over time, or/and are not present in all
the population. To fill these gaps, a use of multimodal biometrics is a first choice
solution [1].
This paper describes a multibiometrics method for human recognition based on
new teacher vector identified as spectrum eigenface, and spectrum eigenpalm. The
proposed combination scheme exploits parallel mode capabilities of the fusion feature
vectors in matching level and invokes certain normalization techniques that increase its
robustness to variations in geometry and illumination for face and palmprint. The
correlation distance is used as a similarity measure. A threshold value is used to
prevent the imposter for being recognized. Experimental results demonstrate the
effectiveness of the new method compared to the unimodal biometrics for spectrum
eigenface/eigenpalm.
Keywords
)
