Design And Implementation Of Embedded Secure Palmprint Recognization System

With the development of society and scientific technology, higher requirements are made for security of information and systems, and the biometric identification and authentication technology is an effective method of security enhancement for information and systems. Palmprint recognition technology is gaining more and more attention recently, due to its advantages such as easy capturing, abundant image information, high recognition accuracy and difficulty in faking. The thesis mainly investigates in the algorithms of various stages of palmprint recognition, and the embedded security palmprint identification system with a dynamic password module is designed and implemented in this paper.The main contents and major result of research in the thesis are listed as follows:1,The market demand, functional requirements and performance requirements for embedded security palmprint recognition system are analyzed in detail. And the overall design of system is introduced from the sides of development platform, the functional structure and workflow.2,In the stage of palmprint orientation and segmentation, a novel algorithm for palmprint orientation and segmentation called Shift and Image Subtraction Method is presented in the thesis after detailed analysis for advantages and disadvantages of the existing palmprint orientation and segmentation algorithm. The Shift and Image Subtraction Method has only three times of the image shift operations, three times of image subtraction operations, two times of image and operations, and without the time-consuming multiplication and convolution operations, so the computational complexity of algorithm is reduced and the runtime of algorithm is speeded up.3,In the feature extraction stage, according to the characteristics of GABOR wavelet transform, the imaginary part of two-dimensional Gabor wavelet filter with three-level Quantization is selected. And a novel algorithm for palmprint feature extraction called Interval Filtering Feature Extraction Method is presented, which the filtering and sampling is implemented at the same time and only the sampling image data points are filtered with two-dimensional Gabor wavelet. This method speeds up the runtime of the feature extraction stage without reducing the recognition accuracy.4,In the matching recognization stage, an improved method for palmprint matching recognization called Marked Matching and Score-level Fusion Method is presented in the thesis after detailed analysis for disadvantages of the existing FusionCode algorithm. First, the precise marked location is implemented after palmprint feature extraction in each direction. Then, the independent Hamming matching is made according to location mark. Finally, the matching scores in the different directions are fused at matching score level which makes palmprint feature of the different directions will not affect each other.5,A embedded palmprint image acquisition device with compact size and embedded security palmprint identification system with a dynamic password module are designed and implemented. The system mainly contains palmprint image acquisition module, registration of palmprint images module, registration of palmprint features module, registration of personal information module, palmprint detection module, palmprint recognition module, dynamic passwords module, human-computer interaction module, and so on.