Research On Recognition Technology Of Palmprint, Palmvein And Its Fusion

Nowadays, automatic authentication using biometrics is becoming more and more popular. This thesis is focused on the implementation of authentication technology by fusion of palmprint and palmvein. Two novel palmprint recognition algorithms are proposed. In order to industrialize the palmprint recognition technology, some prototypes of automatic palmprint recognition system are researched and established. The algorithm and system design of palmvein recognition are preliminarily investigated .The fusion algorithm of palmprint and palmvein using receptive filed and rattlesnake bimodal cell neurodynamics is also discussed and a certain scale of multimodal biometrics databases are established. The main contributions can be exhibited by the following aspects:1. A palmprint recognition algorithm base on moirépattern is introduced. Moirépattern is a useful means for replicated diffraction gratings and analyzing grating imperfections. A moirépattern occurs whenever two contour maps of similar light and dark contours of roughly equal width are overlaid one on top of the other. It's found that the ridges between the ring finger root and little finger root show a two-dimensional periodic structure with a certain degree of polar-like symmetry. Ridges can be viewed as distorted circular gratings, which motivate the idea of superimposing them in order to obtain moirépatterns. An interested region is cropped from the original palmprint and we identify person by extracting moirépatterns from this region.2. Palmprint recognition methods by scene matching are introduced. The process of palmprint verification and scene matching have several similarities, they all can be viewed as the process of matching one template with another sample taken under different illumination, sensor pose geometry, or type of sensors. In the stage of verification of our approaches, the process of extracting ROI (Region of Interest) is not needed any more; template is directly matched with the test palmprint with the help of the mature scene matching techniques. Whether template and the test palmprint are from the same person or not is decided by their matching score. In the case that the test palmprint and template are from the same person, the template's exact position in the test palmprint can even be located. A set of experiments is performed on PolyU palmprint database and NUDT palmprint database to demonstrate the effectiveness of our approaches.3. Based on years of research, a series of automatic palmprint recognition systems are established. The ultimate objective of palmprint recognition is to design a practical automatic palmprint recognition system. We have designed some palmprint recognition systems based on digital camera, video camera, ARM development board and scanner, which can both collect users'palmprints and identify their ID. Some parameters, such as size, weight, response time, recognition accuracy and the price of these systems can meet the needs of some practical applications.4. With the help of mature fingerprint recognition algorithm and retinal vasculature segmentation method,the preprocessing and feature extraction of palmvein are preliminarily investigated. The focus of image fusion is placed on the application of neurodynamics of vision models, and the characteristics of the bimodal cells of rattlesnakes are researched. Fusion of palmprint and palmvein is implemented by adopting the receptive field, i. e. ON-center OFF-surround and OFF-center ON-surround opponent property structure. An image acquisition device which can collect palmprint and palmvein images in the same position and at the same time is deviced, and a small palmprint and palmvein database is built by this device. Some experiments have been done to evaluate this fusion method on the palmprint and palmvein database.5. A certain scale multi-model biometric database is established. The participants are chosen from the students and teachers of our university; the multi-model biometric database is established by the existing fingerprint, face, Iris, palmprint and palmvein capture devices. We hope that the scale of the database can keep expanding, and the database can be open and used for international and domestic counterparts.