Amidst the numerous biometric modalities, the hand vein recognition has been famous for its high accuracy and stability. The idea of our proposed method is to extract new features from the hand vein, such as; vein direction, its length and combined veins that are considered to be a unique feature for the hand vein of each particular person. This images exposed to filtering techniques, in addition to using enhancement approaches and segmentation processes. This was followed by dividing this study into two parts: the first part used a traditional Pixel by Pixel method to test the processed images. On the other hand, the second part (which is the proposed method of this study) used the “Modified Hough transformation”, which is the novelty of the proposed method, to extract the structural features; vein lengths and its angles based on the studied images. Fusion between these structural features and brightness indicator (white pixels) has been followed to be considered as new features for vein image then to be classified. The brightness indicator has been calculated based on the study of number of white pixels before and after Hough transformation. These techniques were applied to a total number of 600 images collected from 100 individuals belonging to a diverse demography of age groups. Finally, matching experiments were implemented for both parts, and the results obtained revealed that the second part yielded 99.5% accuracy compared to 98.5% reaching by traditional method. |
