摘 要
人脸识别技术是模式是别和计算机视觉研究中的一个重要领域,在边防安全、视频监控、身份验证等方面有重要的应用价值。人脸检测是快速、准确识别人脸的前提,其目的是将人脸从图像背景中检测出来。传统的课堂签到费时、费力、效率低。提出基于人脸识别的课堂签到系统,让课堂签到更加高效化、规范化、便捷化。
本文以python为开发语言,基于tensorflow平台,使用django +celery 作为主体框架 ,前端使用 opencv.js +html 实时检测摄像头人脸 ,后端使用django 获得人脸检测任务 发送给celery 做特征提取和人脸分类,数据库采用SQLite 。人脸识别主要使用 dlib做人脸提取,使用 resnet 做人脸特征提取,然后使用计算欧氏距离分类出人脸,从而实现人脸识别。人脸识别系统可以进行注册、签到、用户管理等。系统界面简洁、识别迅速、使用方便。
论文首先介绍了人脸识别系统的现状及开发背景,然后论述了系统的设计目标,系统需求和总体设计方案,较详细的论述了系统的详细设计和实现,最后对人脸识别系统进行了一些具体测试。
Abstract
Face recognition technology is an important field of pattern and computer vision research, and has important application value in border security, video surveillance, identity verification and other aspects. Face detection is the premise of fast and accurate face recognition, its purpose is to detect the face from the image background. Traditional classroom attendance is time - consuming, laborious and inefficient. A class check-in system based on face recognition is proposed to make class check-in more efficient, standardized and convenient.
This paper uses Python as the development language, based on TensorFlow platform, and uses Django + Celery as the main framework. OpenCV.js + HTML is used in the front end to detect the camera face in real time, Django is used in the back end to obtain the face detection task and send it to Celery for feature extraction and face classification, and SQLite is used in the database. Face recognition mainly uses dlib to do face extraction, the use of resnet to do face feature extraction, and then use the calculation of Euclidean distance classification of faces, so as to achieve face recognition. Face recognition system can be registered, sign-in, user management, etc. System interface is simple, rapid recognition, easy to use.
The paper first introduces the current situation and development background of face recognition system, and then discusses the design objectives, system requirements and overall design scheme of the system, a more detailed discussion of the detailed design and implementation of the system, finally on the face recognition system for some specific tests.
目 录