Joint Transform Correlator Fingerprint Identification System Design And Realization

Fingerprint recognition technology is a means to achieve identification through a computer. It has a very important role of biometric technology. Fingerprint recognition technology has been widely used in various fields, which means that it can affect the daily life if millions of people. Fingerprint recognition technology can prevent unauthorized access by replacing the personal identification code and password. There will be a new era of personal authentication thanks to the rapid development and wide application of fingerprint recognition technology. But in the algorithm implementation of existing fingerprint recognition technology, the accuracy and real-time of the system is difficult to coordinate, and most of the high accuracy of the algorithm will bring the computational complexity and impact of real-time. In response to this situation, this thesis proposes to achieve fast and accurate fingerprint recognition based on JTC as the core system.The main content of this thesis are as follows. Firstly, preprocess the collected fingerprint. This thesis introduces the concept of a wavelet transform filtering and direction filtering on the basis of the traditional fingerprint preprocessing, which makes the fingerprints after the pretreatment is more suitable as input and improve the reliability of the system. Then, on the basis of accuracy of computer simulation system, this thesis uses two programs. The first is based on the optical bench by selecting the appropriate parameters of the optical device to set up JTC. The second is CUDA-based high-speed computing simulation system by programming to enable the GPU instead of CPU to process the major computational.The final results show that, using fingerprint recognition technology based on optical platform after a series of fingerprint preprocessing, we can achieve more significant cross-correlation peak by JTC, and the operation processing in the speed of light. The cross-correlation peak got by using CUDA integrated computer simulation technology is also more obvious and achieve the desired effect and feasibility. In this thesis, the computer simulation only use part of the GPU-accelerated algorithm. Therefore, by improving the algorithm, the system computing speed will be greatly improved. I hope to further strengthen the two system’s utility and real-time.