粒子群算法优化BP神经网络的指纹识别技术  

Fingerprint Identification Technology of BP Neural Network Based on Particle Swarm Optimization

作　　者：马少华[1] 曹三民[1] 

机构地区：[1]沈阳建筑大学信息与控制工程学院,辽宁沈阳110168

出　　处：《科技广场》2011年第7期40-42,共3页Science Mosaic

摘　　要：指纹识别技术是当今应用最广泛的生物识别技术之一。在指纹识别过程中,图像处理、特征提取、匹配等过程数据量庞大,计算比较烦琐。BP神经网络具有良好的自学习能力、强大的分类能力和容错能力,将其应用到指纹识别中是可行的。为改进BP神经网络计算速度较慢,梯度下降法不能处理一些不可微传递函数的问题,采用粒子群算法对BP算法进行优化,提高了指纹识别的速度和准确度。Fingerprint identification technology is one of the most widely used biological recognition technology today. In the process of fingerprint identification, image processing, feature extraction, matching process have large amounts of data to be addressed and the calculation is also very troublesome. The BP neural network has good self-learning ability, strong classification ability and fault tolerance and it is feasible to be used in fingerprint identification. Meanwhile, the BP neural network also contains some problems such as slow computing speed, the gradient descent method can＇t deal with non-differential transfer function. This paper adopts the particle swarm optimization to optimize BP algorithm and improves the speed and accuracy of fingerprint recognition.

关 键 词：指纹识别 BP神经网络 粒子群算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]