基于卷积神经网络的人脸识别系统的设计(Python)(论文10000字,外文翻译,参考代码,流程图,人脸图像库)

摘要：随着社会的进步与发展，个人信息的保护变得十分重要。传统的密码保护方式已经不再满足人们的需要，为此，进化出了多种多样，基于个人独特特征的密码形式，如指纹识别系统与人脸识别系统。而人脸识别系统因为具有仿造性低，信息采集方便等多种优势，成为了继指纹识别之后，大热的一项密码形式。传统的神经网络已经满足不了现有的图像识别所需的计算能力。而卷积神经网络则完美解决了这一问题。本文提出，应用卷积神经网络于人脸识别系统中。卷积神经网络启发于视觉系统的结构，其人工神经元能够只对部分信息进行响应，再将响应传输到下一层，这样能够大幅的减少计算量的同时保证极高的正确率，从算法方面研究卷积神经网络在人脸识别系统中的应用与优化，提高识别精确性与效率。

关键词：卷积神经网络，人脸识别，图像识别，深度学习

Face Recognition Based on Convolutional Neural Network

Abstract：With the progress and development of society, the protection of personal information becomes very important. Traditional password protection methods no longer meet people's needs. To this end, a variety of cryptographic forms based on individual unique characteristics, such as fingerprint recognition systems and face recognition systems, have evolved. The face recognition system has become a password form after the fingerprint recognition because of its low imitation and convenient information collection. Traditional neural networks have not been able to meet the computing power required for existing image recognition. The convolutional neural network solves this problem perfectly. The convolutional neural network is inspired by the structure of the visual system. His artificial neurons can respond to only part of the information and then transmit the response to the next layer, which can greatly reduce the amount of calculation while ensuring a very high accuracy. This paper proposes to apply convolutional neural network to face recognition system to study the application and optimization of convolutional neural network in face recognition system from the aspect of algorithm, and improve recognition accuracy and efficiency.

Key words：Convolutional Neural Network, Face Recognition, Image Recognition, Deep Learning

目录

1 绪论    1

1.1 卷积神经网络(CNN)的发展    1

1.2人脸识别的发展    3

1.3 人脸识别的应用领域    5

1.4 本章小结    7

2 卷积神经网络的原理及其架构    8

2.1 卷积神经网络的原理    8

2.2 卷积神经网络的架构    9

2.3 几个经典的CNN网络分析    12

2.4 本章小结    14

3 卷积神经网络在人脸识别系统中的应用    15

3.1 面部识别的流程分析    15

3.2图像预处理    16

3.3建构卷积神经网络    16

4实验结果    20

结论    23

参考文献    24

致谢    25