| Biometric personal identification is an important research area aiming at automatic identity recognition and is receiving growing interest from both academia and industry. There are two types of biometric features: physiological (e.g.face, iris pattern and fingerprint) and behavioral (e.g. voice and handwriting). Handwriting-based writer identification is a behavioral biometric identification approach. Handwriting is easy to obtain and different people have different handwritings. Handwriting based personal identification has a wide variety of potential applications from security to financial activities.Handwriting-based writer identification is an active research topic in computer vision and pattern recognition since last century. However, there are still many unsolved problems existing in this field. And all works in this field still stay in lab .The main purpose of this research is: presenting new methods and algorithms based on deep analysis on all known technologies in this field. Further, developing a software system for handwriting-based writer identification and putting it into business practice.This research owns good value on society and economy. First, this research will develop a mature software system used for handwriting-based writer identification, which is first one in China. In this couse, we must overcome a series of problems from academic theory to practical application, as will greatly impel the development of handwrting-based writer identifcation. Also, this software system has a wide application in police departments and will bring great benefit for society and ecomomy. Besides those, technologies used in this software system can also be used in other fields such as finance field, communication field and so on. Last, this software is also a sign which means automatic handwriting-based writer identification comes to maturity and can be used in industry.Chapter 1 is introdcution of handwriting-based writer identification, in which object and purpose of this research are presented.In chapter 2, we introduce the process of manual writer identifcation, find out advantages of manual process and automatic process, and point out the importance of human computer interaction(HlC). Then, we present a whole flow chart of automatic handwriting-based system.In chapter 3, a series of pre-process algorithms are presented, which make good efforts.In chapter 4, according to human's feel on object's shape, a algorithm based on momentfor writer identifcation is presented, of which the matching function is modificatory block distance. Aiming at large dynastic range of various moments, we give a normalization method in this chapter.Chapter 5 mainly discusses High-order-related algorithm used in writer identification. High-order-related algorithm uses correlativity of strokes to identify document writer, and has good results in experiments. However, the algorithm needs high computational requirements. In this chapter, we presents a new high-order-related algorithm which keep high accuracy and significantly reduces computational requirements.Texture analysis is a hot approach in automatic handwriting-based writer identification. In chaper 6, basic knowledge and common methods are introduced and a algorith using Gabor filter to extract texture features is presented. Further, we give a idea based on wavelet for writer identification.Professional examiners' experience are fortune in this field, which are accumulated in a long time and are correct in most cases. In chapter 7, we discuss integration of experts' experience and computer systemm, and then pointed out some faulties of known expert systems. Further, we present a expert system for handwriting-based writer identification, which has creative ideas in feature's description and matcting.Chatper 8 is introduction of our software system. First, we introduce whole structure and all modules of this system . Then, we discuss relative software technologies, including C ++ Builder and Matlab.In chapter 9, we sum up this research a...
|
PDF Full Text Request