Palmprint recognition system based on principle-lines feature using Euclidean Distance and neural network

Human recognition system can be done based on the characteristics of palmprint. The palmprint identification system is the palmprint matching process tested with the entire palmprints have been registered on the database. The result is to recognize one individual palmprint. This system consists of software completed with a webcam as input to generate input images. In the system, feature extraction uses detection of the principal-line features and then the image is divided into blocks and form a feature vector of palmprint. The process of feature extraction starts with the palmprint lines acquisition which plays an important role towards the success of recognition. The system developed was tested using 90 palmprint images from 30 individual persons with three samples of palmprint were acquired from one person. Two of three palmprint samples were used as test images, while the rest is used as the reference image. From the test results using Euclidean Distance as its similarity measure, it can be concluded that the recognition system based on principal lines feature of palmprint is well-performed with the successful recognition of the palms tested reaches 100%, both for testing using palm images which have been trained as well as for testing using outer images. While, for recognition using neural network, the results are: The test results show that the performance of palmprint recognizing system with 3 positions of test images provides success rate 88.88 %. The test with straight vertical position provides success rate 90 %. The test with 90° angle to the right position provides success rate 93.33 %. The test with 90° angle to the left position provides success rate 90 %.