Summary: | 碩士 === 淡江大學 === 電機工程學系碩士班 === 93 === The correct minutiae extraction is very important in an automatic fingerprint identification system. However, the presence of noise in poor-quality images will cause many extraction faults, such as the dropping of true minutiae and inclusion of false minutiae. Nowadays, most fingerprint identification systems are based on precise mathematical models, but they can not handle such faults properly. As we know, human beings are good at recognizing fingerprint pattern. Therefore, a human-like method is applied. This paper presents an adaptive fuzzy logic and neural network method which is fast and has variable fault tolerance. We implement a fast fingerprint database system with fault tolerance. Before neural network training, every fingerprint is encoded by a fuzzy image encoder. Then the result of training is saved in a database. The training time is 3 seconds. The matching time is 0.08 second. When the threshold is 0.9, the FAR is 0% and FRR is 0.23%. Our experimental results have shown that this fingerprint identification method is robust, reliable and rapid.
|