Research On The Key Technologies For Automatic Fingerprint Identification System

As a significant approach to personal identification, biometrics is attracting unprecedented attention. Fingerprint identification is one of the most representative technologies of biometrics and has achieved great success both in police and civil applications.Automatic fingerprint identification system (AFIS) involves the technologies of image processing, pattern recognition, computer and sensor. It is developed with the increasing demand for fingerprint identification and the origins of the research can be traced back to the early 1960s. With a half century of development, the performance of AFIS has been greatly improved. However, due to the complexity and uncertainty of fingerprint identification, there are still many problems to be resolved for further improvement of AFIS.The dissertation primarily investigated the fingerprint identification algorithm which is one of the most key technologies of AFIS, and further constructed a practical system based on the research of algorithm. The main contributions of the dissertation are as following:1. The dissertation investigated fingerprint image quality evaluation and proposed a new evaluation method considering the limitations of the existed methods. The proposed method makes comprehensive analysis on various features of a fingerprint image by utilizing the fuzzy inference system (FIS), and carries out accurate image segmentation by utilizing the support vector machine (SVM). Based on this, the global quality of fingerprint image is finally evaluated. Experimental results show that the proposed method has the ability to distinguish between fingerprint images with various qualities and may contribute to promote the performance of AFIS.2. The dissertation made a survey on fingerprint image enhancement and proposed a novel enhancement algorithm which adopts Log-Gabor filters to enhance fingerprint images for the first time. The proposed algorithm effectively overcomes the drawbacks result from the limitations of traditional Gabor filters in virtue of the unique frequency response characteristic of Log-Gabor filters. In this algorithm, a bank of Log-Gabor filters is constructed according to the texture features of fingerprint image, and the image is enhanced by frequency filtering. Experimental results show that the proposed algorithm can effectively improve the quality of fingerprint images and help to ensure the accuracy and robustness of AFIS.3. The dissertation investigated fingerprint matching and presented a new matching method based on the analysis and comparison of the existed methods. This method firstly searches the minutiae pairs which satisfied the requirement of local matching by utilizing the ridge information. Then, the minutiae sets are aligned globally through coordinate transform, and the alignment parameters are optimized by using affine transform and least square method. Finally, minutiae matching and directional field matching are respectively achieved based on the minutiae alignment, and the two matching scores are fused into a single to obtain a more reasonable result. Experimental results show that this method contributes to ensure the accuracy of fingerprint identification as well as possesses the real-time performance.4. The dissertation designed and realized an embedded fingerprint identification system which owns independent intellectual property. In this work, both the hardware platform and the software system are investigated and designed, and the algorithm of fingerprint identification is also successfully transplanted into the embedded system. The hardware structure has been awarded the Chinese Patent and the corresponding software system already passed the software product register test of Chongqing software test center. The achieved technical performance level shows that this system is of considerable practical value.The dissertation systemically and deeply investigated the theory and methods of AFIS, brought forth new ideas on fingerprint image quality evaluation, fingerprint image enhancement and fingerprint matching, and eventually realized an embedded fingerprint identification system with independent intellectual property. This study may provide theoretical and technical support for further developing the AFIS with higher performance. It is of active significance to promote the development and application of automatic fingerprint identification technology.