Segmenting Finger Inner Surface for the Purpose of Human Recognition

Patterns of Finger Inner Surfaces (FISs) can provide high security contributions, due to the fact that they have unique references and they completely vary from person to person. In this paper, a FIS has been exploited in the field of pattern verification. Firstly, the FIS is segmented by adopting an innovation technique by utilizing multiple image processing methods. Secondly, a feature extraction method based on the Wavelet Discrete Transform (DWT) and Coefficient of Variance (CV) is applied. Thirdly, an intelligent software is implemented to confirm the ability of the FIS is employed, where Neural Network (NN) performances are applied in this investigation. Random codes are used for the segmented FISs of each hand of individual in terms of verification. In addition, the FIS patterns for the right and left hands have been examined using CASIA multi-spectral palm images. As a result of this research, the False Acceptance Rate (FAR) and False Rejection Rate (FRR) obtained (0.3333%) and (0.8889%), respectively.