【作者】 陈香华；

【导师】 叶吉祥；

【作者基本信息】 长沙理工大学 ， 计算机应用技术， 2006， 硕士

【摘要】 在网络技术迅猛发展的时代,网络应用、电子商务的发展使我们越来越多地依赖智能卡、身份证、密码、数字证书等安全措施进行身份识别和安全认证。但各种措施都或多或少地伴随着要携带、易遗忘、易丢失、被盗用等不方便和不安全的因素。随着各种安全认证措施相继在网络应用上搁浅,生物识别技术开始进入安全认证的舞台。生物识别技术是依据人的生物特征的惟一性、不变性和可测量性进行个体识别的。指纹识别技术以方便易用、高准确率和低价格等诸多优势倍受关注,近年来有了长足的进步和发展,是目前人体特征识别技术市场占有率最高的方式之一,也是目前生物识别技术发展中最成熟的。虽然目前有很多指纹系统的商业产品和研究成果,但仍不能达到某些特定应用的性能要求。本文在原有指纹识别技术的基础上,融入一新的图像处理技术,打破了常规的指纹特征描述方法。纹理图像的个性特征主要表现在图像像素之间的空间几何特征上,利用脉冲耦合神经网络可以提取纹理图像的点火时间序列信号,这是从一个新的角度描述指纹特征。经过对捕获率序列的仔细研究发现,它不仅反映了局部图像的灰度分布信息,更重要的是它还包含了图像中相邻象素之间的空间几何信息。因此,在所提取的特征中,纹理的几何特性得到了充分的体现。更多还原

【Abstract】 We more and more depend on various security measures with the development of the network application and electron business affairs ,such as smart cards, identity cards, password, digital certificate etc, to carry out identity recognition and safety authentication when the network technology develops rapidly. But all of that are usually discommodious and insecure to some extent, for example toting, forgetting, losing and peculating and so on. These disadvantages impact the use of safety authentication measures in the network application, so biometric technology comes into it synchronously. Biometric technology lies on the particularity, the invariability and the measurability of biometric characteristic to recognize the individual. Fingerprint recognition technique is used widely because of its easy use and high accuracy, and it makes great progress and develops adequately in recent years. At present, it is one of recognition technique which owns the highest occupancy ratio in the market of biometric technology and it is also the ripest of them. Although Fingerprint recognition system possesses many commerce products and research productions, it still can’t achieve special demands. This paper presents a new way of image processing in the old foundation to break the routine way of describing the characteristic of fingerprint.The especial characteristic of texture image mainly lies on the space-geometry -characteristic of pixels of image. It is a new way of describing fingerprint image’s characteristic that using PCNN obtains the ignition-time-sequence-signal’s characteristic. After researching, we can find that the capture-ratio sequence not only reflects local image’s grey distribution but also includes the space-geometry’s characteristic of neighboring pixels of image, so the space-geometry-characteristic of texture in the obtained characteristics are reflected further.更多还原