摘要人脸录入系统在当今社会中有着广泛的应用，比如一些门禁系统、身份识别仪、考勤系统等等。在人脸录入过程中，最重要的一步就是人脸识别。实际应用中，先要进行图片判断，看其中是否有人脸，这就是人脸检测。然后对人脸图像进行质量评价。图像质量评价是通过设计算法，给出和人的主观感受相一致的评价。60980

本课题是通过对目前流行的图像质量评价算法的学习，归纳出一种综合评价算法。选择图像中人脸位置、姿态、对比度、清晰度和明亮度作为评价指标，并采用Visual C++ 2008和OpenCV实现人脸图像质量评价，最后将综合分值用直方图表现出来。在人脸检测过程中，还需要判断人眼是否水平，保证录入图像人眼水平。本课题在实际应用中具有一定的研究价值。

毕业论文关键字：图像质量评价，人脸识别，综合评价，人脸图像

 毕业设计说明书（论文）外文摘要

Title    Face-entry System Based On Image Quality Assessment                 

Abstract

Face-entry system has been widely applied in today’s society,such as some access control systems,instrument identification,time and attendance systems,etc.In the face entry process,the most important step is recognition.In practical applications,first for image judgments to see if any of them is the face.This is called Face Detection.Then we should evaluate the quality of the face image.Image quality assessment will qive some evaluations consisted people’s subjective feelings by designing algorithm.

This article presents a comprehensive evaluation algorithm based on learning the current popular image quality assessment algotithm.The image location,stance,symmetry, brightness,contrast and definition are choosed as evaluation indexes, and a face image quality assessment system is developed by using Visual C++ 2008 and OpenCV.Finally,the composite scores are showed by Histogram.In the face detection process,we must determine whether the level of the human eyes to make sure that the human eyes of entryed images are level.This topic has a certain research value in practical applications.

Keywords:Image Quality Assessment ;Face Recognition ;Comprehensive Assessment ;Face Picture

目次

1绪论 1

1.1 课题研究意义 1

1.2 人脸识别的现状 1

1.3 研究目标 3

1.4 研究的方法及技术路线 4

1.5 本论文的结构 4

2 人脸检测的常用方法及相关技术 5

2.1人脸检测介绍 5

2.2人脸检测的常用方法 5

2.2.1 AdaBoost人脸检测算法 5

2.2.2 OpenCV介绍 7

3视频人脸检测录入的实现 8

3.1系统环境 8

3.2功能需求分析 8

3.2.1人脸位置评价算法 9

3.2.2人脸姿态评价算法 9

3.2.3图像对比度评价算法 10

3.2.4图像清晰度评价算法 11

3.2.5图像明亮度评价算法 12

3.2.6人脸图像质量综合评价算法 12

3.3部分程序代码