Vein palm recognition model using fusion of features

Abstract

One of the most promising mechanisms in the field of security and information safety is authentication based on palm vein. The main reasons that vein palm becomes an authentication method is because of its distinctive privacy, as it is difficult to manipulate or change its results, because of the location of the vein within the palm. With the use of this technology, it has become easy to maintain data from unauthorized access and unwanted persons. In this work proposed model are suggested that contain four stages to reach the results: in the first stage is the pre-processing stage where histogram equation was used to enhance the image and the properties are shown, the second stage is the extracting the properties where, Gabor filter and 2-discrete wavelet filters are suggested for features extraction, where it is considered one of the most important filters used to extract the features, as well as in the third stage "PCA" are used for data or features reduction, because of its advantages in analyzing the features and reducing the spacing between them. As for the last stage, the Euclidean distance used to measure the spacing. The results were acceptable and convincing, since the similarity ratio 96.2%. These results were obtained after several tests and using the Gabor filter with 2D-discrete wavelet transform and principal component analysis (PCA), I got the best results.