基于神经网络的指纹识别答题.doc

西南交通大学本科毕业设计（论文） 第 PAGE II页 西北民族大学 2011级毕业论文 基于神经网络的指纹识别 年 级: 2011级 学 号: 姓 名: 专 业: 自动化 指导老师: 二零一五年五月 西北民族大学本科毕业设计 第 PAGE \* MERGEFORMAT 20页 共 NUMPAGES \* MERGEFORMAT 26页 西南交通大学本科毕业设计（论文） 第 PAGE I页 摘 要 随着科学技术的发展，信息安全逐渐被人们所重视，指纹识别技术因其唯一性和安全性逐渐进入了人们的视野，并广泛应用于刑事侦查、设备安全和个人信息等诸多领域。在本文中对指纹图像的必要步骤进行了分析，对指纹图像的预处理，剔除指纹图像中可能存在的伪特征点，对指纹图像的特征提取，将提取的指纹特征训练神经网络，利用神经网络技术对指纹的特征进行识别。本文将BP神经网络和RBF神经网络对指纹识别的结果进行对比，比较两种方法在指纹识别上的性能，利用Matlab软件对该方法进行了实验验证。结果表明RBF神经网络在指纹识别方面比BP神经网络拥有更好的泛化性能。 关键词：指纹识别； 特征提取； 神经网络 Abstract The information security is paid great attention to people with the development of science and technology, and the fingerprint recognition technology entered peoples vision gradually with it’s uniqueness, and security.It is widely used in criminal investigations, equipment safety and personal information, and so on. In this paper, the various steps of fingerprint recognition are: pre processing of fingerprint images, the possible existence of burr or pseudo feature points are eliminated in the fingerprint image, the fingerprint image rotation and translation after the feature extraction, fingerprint feature extraction as the input vector of neural network to identify the input vector, using the technology of neural network. This paper compares the results of BP neural network and RBF neural network for fingerprint recognition, a performance comparison of two methods in fingerprint recognition, the method is verified by using Matlab software. The results show that the RBF neural network in the field of fingerprint recognition has better performance than BP neural network. Keywords：fingerprint identification; feature extraction; neural network 目 录  TOC \o 1-3 \h \z \u  HYPERLINK \l _Toc419630166 摘 要  PAGEREF _Toc419630166 \h II  HYPERLINK \l _Toc419630167 Abstract  PAGEREF _Toc419630167 \h III