论文总字数：36660字

摘 要

现代社会中，随着科技和信息技术的的不断发展，建立科学并且规范的教学管理体系，重中之重，这样确保了学校的快速发展。在高校学生课堂学习状态分析与评价系统设计中可以解决分析学生上课不认真而影响的教学质量，对高校课堂中的学生进行分析与评价，及时给教师反馈信息。人的喜怒哀乐都会体现在表清上，通过人脸识别、模式识别、人工智能、图像处理等这些快速发展的技术在交通安全、医疗、人机交互、医疗等领域发挥着广泛的应用，有着重要的研究价值，用这些技术来提升教学质量，为本次研究带来新的想法和处理方式。

本次研究的主要工作如下：

1. 将高校学生课堂学习的状态和表情状态结合起来，通过剖析学生们的各种状态，联系实际情况的表情识别，心理学，行为学，对学生在课堂中的专注度进行划分，并验证合理性，达到符合实际。
2. 运用开源计算机视觉库（OpenCV）为这整体补充新的功能的历程和算式的集合，通过前端获取得到人脸的头像，接着使用哈尔特征(Haar-like features)的算法，找到人脸并且把人脸图片切出来，运用Dlib库定位人脸，再通过多个任务训练的不同模板的卷积神经网络（Convolutional Neural Networks, CNN）模型，把人像图片转达给锻炼好的网络去判断和预测表情。对人脸对象检测、人脸特征提取、分类，增强图像的对比度，并且应用基于表情识别的综合深度学习，由此完成对人脸的检测和定位。
3. 然后应用Python语言开发技术，选取PyQt5去实行GUI界面的开发基于Pyton和Qt库中的API，设计并开发了基于人脸识别的高校学生课堂学习状态分析与评价可视化系统。该系统可以把学生的上课状态以个性化和可视化的方式呈现出来，方便教师去进行分析和管理班上的学生。

该论文提出了通过人脸识别和卷积神经对人脸的检验和定格等算式和技术，探究高校学生在讲堂中的状况区别和评价，设计并实行出源于人脸识别的学生状态分析系统。应用高端的现代技术在教学过程中便于管理学生，提高教学质量。本次系统通过测试，准确率可以达到90%，系统可以用于分析学生课堂状态，给教师反馈学生的学习状态，让课堂环境和教师教学环境更加适合学习和教学。虽然本次系统的准确率可以达到80%，但还是需要在深度学习和界面优化做进一步的改进和提高。让学校的教学环境更加的生动有趣，提高学生们的学习兴趣，让高端的技术走进学校，改变环境。

关键词：表情识别；Python；CNN；深度学习；OpenCV；Haar

Analysis and Evaluation System of Classroom Learning Status of College Students

Abstract

In modern society , with the continuous development of technology and information technology, the establishment of scientific and standardized teaching management system, thus ensures the rapid development of the school.In the design of the system of analysis and evaluation of students' classroom learning status, we can solve the problem of analyzing the teaching quality that students are not serious in class, analyze and evaluate the students in the classroom, and give feedback information to teachers in time.People's joys and sorrows will be reflected in the table, through face recognition, pattern recognition, artificial intelligence, image processing and other rapid development technologies in traffic safety, medical, human-computer interaction, medical and other fields play a wide range of applications, has important research value, using these technologies to improve the quality of teaching, for this study to bring new ideas and processing methods.

The main work of this study is as follows:

1. By analyzing the students' various expression states, combining with the actual situation of expression recognition, psychology, behavior, the students' concentration in the classroom is divided, and the rationality is verified. To meet the actual situation.
2. Using open source computer vision library (OpenCV) library to add new functions to the whole library, the face image is obtained by front-end, and then the Halt sign (Haar-like features) algorithm is used to find the face and cut the face picture out. The Convolutional Neural Networks, CNN model of convolution neural network with different templates trained by multiple tasks is used to judge and predict facial expressions. Face object detection, face feature extraction, classification, enhancement of image contrast, and application of comprehensive depth learning based on expression recognition to complete face detection and location.
3. Using Python language development technology, using the program PYQT5 realize the GUI interface to develop the API, design based on the face recognition and the visual system of classroom learning state analysis and evaluation of college students. The system can present the students' class status in a personalized and visual way, which is convenient for teachers to analyze and manage the students in the class.

This paper presents the detection and positioning of face recognition and deep learning, studying and discussing the state division and evaluation of college students in class , and designing the classroom state analysis system based on expression recognition, detecting and positioning the students' concentration in the teaching process.The application of high-end modern technology in the teaching process is convenient to manage students and improve teaching quality.The system through the test, the accuracy rate can reach 90 %, the system can be used to analyze students' classroom state, give teachers feedback to students' learning state , so that the classroom environment and teachers' teaching environment are more suitable for learning and teaching .Although the system accuracy that we do can reach 80 %, further improvement and improvement in deep learning and interface optimization are required .Make the school's teaching environment more vivid and interesting, improve students' interest in learning, let high - end technology enter the school, change the environment .

Keywords : expression recognition ; Python;CNN; deep learning;OpenCV;Haar

目录

高校学生课堂学习状态分析与评价系统 3

摘要 3

Analysis and Evaluation System of Classroom Learning Status of College Students 4

Abstract 4

目录 6

第一章绪论 8

1.1研究背景与意义 8

1.2研究现状 8

1.3主要工作内容 9

1.4论文组织结构 10

第二章相关技术基础 11

2.1图像处理 11

2.1.1灰度化 11

2.1.2高斯滤波、均值滤波、中值滤波、双边滤波（去噪） 11

2.1.3直方图均衡化 12

2.1.4亮度、饱和度调节 12

2.2图像识别 13

2.3神经网络与CNN 13

2.4 Python简介 14

2.5OpenCV 14

2.6本章小结 16

第三章系统需求分析 17

3.1前期分析 17

3.2系统功能需求分析 17

3.2.1功能需求分析 17

3.2.2功能需求用例图 18

3.3系统非功能需求分析 18

3.3.1系统性能分析 18

3.4本章小结 19

第四章系统设计 20

4.1系统总体设计 20

4.1.1系统分层结构设计 20

4.1.2系统功能模块设计 21

4.1.3系统总体工作流程设计 23

4.2系统详细设计 25

4.2.1控制模块的设计 25

4.2.2表情识别模块的设计 26

4.2.3评价模块设计 27

4.3核心算法设计 28

4.3.1 OpenCv算法 28

剩余内容已隐藏，请支付后下载全文，论文总字数：36660字