Development Of Digital Image Processing And Recognition System

With the development of the science and technology, and with the exploitation of the field in computer application, a brand-new image processing method has emerged as the times require. That is digital image processing technology. With computer devices, this technology can change images to digital information for storing, processing, transforming and redisplay.The technology of digital image recognition is an important research direction extending from the technology of digital image processing. At present, the application of digital image processing and recognition becomes wider and wider. But as the present level, the ability of computer in outside apperception is still very weak, so it needs more manpower and more material resource to do research on the theories and applications of digital image processing and recognition.Entering upon the "money recognition" and "face recognition" which are the hot points in digital image recognition, the digital image processing and recognition system is studied and discussed.In this paper, what compose the digital image processing and recognition system are introduced firstly. At Chapter 2 to Chapter 4, three parts of this system are introduced, which are "Essential methods of digital image processing", "Face recognition" and "Old paper money recognition". In Chapter 2, the author summaries the essential methods used in this system, which includes two parts: image enhancement and image detection.The method of "approximate neural network with sparse RAM" is used in Chapter 3 for "Face recognition". Author makes an analysis of recognition rate on the effect of the different NN parameters. Visual C++ is used as the programming language to implement this method.In Chapter 4, the edge detection, Fisher linear criterion and neural network are used for "Old paper money recognition". The edge detection is stable in performance and with a good recognition rate, but needs to differentiate the value and sides of the money image; the Fisher criterion needs to differentiate the money image, but its recognition rate is affected by the choice of samples; the NN will get high recognition rate if it differentiates the money's value and sides and low recognition rate if it doesn't.At last, a summary is given to analyse the performance of each part of the digital image recognition system.