For the past two years at the National Laboratory of Pattern Recognition, the author has been working on biometric personal identification. His work includes writer identification, iris recognition and font recognition as described in this Master thesis. The thesis is organized as follows. Chapter 1 gives an introduction of biometric personal identification and an overview of the biometric technology. A number of biometrics-based technologies are briefly introduced and their comparisons are presented. Chapter 2 deals with the structure of a biometrics-based personal identification system. It also addresses the concept and application of multi-biometric personal identification systems. Chapter 3 describes a content independent algorithm for off-line writer identification. The new algorithm takes the handwriting image as an image containing some special texture, and regards writer identification as texture identification. We apply the well-established 2-D Gabor filtering technique to extract features of such textures and a weighted Euclidean distance classifier to fulfill the identification task. Experiments are made using Chinese handwritings from 17 different people and very promising results were achieved. Chapter 4 presents a new algorithm for personal identification based on iris patterns. It is composed of iris image acquisition, image preprocessing, feature extraction and classifier design. The algorithm for iris feature extraction is based on texture analysis using multi-channel Gabor filtering and wavelet transform. Compared with existing methods, our method employs the rich 2-D information of the iris and is translation, rotation, and scale invariant. Chapter 5 describes a new texture analysis based approach towards font recognition. In this new method, we take the document as an image containing some special textures, and font recognition as texture identification. The method is content independent and involves no local feature analysis. Experiments are made using 14,000 samples of 24 frequently used Chinese fonts (6 typefaces combined with 4 styles) as well as 32 frequently used English fonts (8 typefaces combined with 4 styles). Very promising results are obtained.
修改评论