人脸识别技术开源代码及最新趋势汇总

第一部分:人脸识别开源项目汇总

1、Face Recognition

简介:Recognize and manipulate faces from Python or from the command line with the world's simplest face recognition library.

源码网址:https://github.com/ageitgey/face_recognition

案例(Python Code Examples)All the examples are available here.

2、real_time_face_recognition

功能简介:This is a real time face detection and recognition project base on opencv/tensorflow/mtcnn/facenet. Chinese version of description is here .Face detection is based on MTCNN.Face embedding is based on Facenet.

源码网址:https://github.com/shanren7/real_time_face_recognition

 

3、facenet

功能简介:This is a TensorFlow implementation of the face recognizer described in the paper "FaceNet: A Unified Embedding for Face Recognition and Clustering". The project also uses ideas from the paper "A Discriminative Feature Learning Approach for Deep Face Recognition" as well as the paper "Deep Face Recognition" from the Visual Geometry Group at Oxford.

源码网址:https://github.com/davidsandberg/facenet

 

4、face_recognition

简介:使用李子青团队的webface人脸数据集,根据汤晓欧团队的DeepID网络,通过Caffe训练出模型参数,经过LFW二分类得到人脸识别准确率。

  • 根据http://vis-www.cs.umass.edu/lfw/提供的数据集以及pairs.txt得到需要检验的6000对图像(3000对相同人脸,3000对不同人脸)
  • 分别将对应的两个图像分别作为训练好的模型输入,得到两个160维的特征向量。6000对图像依次进行操作,共得到6000对160维特征向量。
  • 计算对应的特征向量的余弦距离(或欧式距离等其他距离),6000对图像依次进行该操作,得到6000个余弦距离(或欧式距离等其他距离),通过选择阈值得到人脸识别的准确率。

源码网址:https://github.com/hqli/face_recognition

 

5、DeepFace

简介:基于开源框架实现的人脸识别、脸脸检测、人脸关键点检测等任务 各个任务分别在FaceDetection, FaceAlignment, FaceRecognition 三个文件中。人脸检测 baseline: 基于基于滑动窗口的人脸检测,将训练好了的网络改为全卷积网络,然后利用全卷积网络对于任意大小的图像输入,进行获取输出HeapMap。

源码网址:https://github.com/RiweiChen/DeepFace

 

6、facerecognition_guide

简介:This is my guide to face recognition with OpenCV2 C++ and GNU Octave/MATLAB. If you research on face recognition, you'll soon notice there's a gigantic number of publications, but source code is very sparse. So this guide is here to change that. Two algorithms are explained and implemented with GNU Octave/MATLAB and OpenCV2 C++ namely Eigenfaces and Fisherfaces.

源码网址:https://github.com/bytefish/facerecognition_guide

Blog网址:http://bytefish.de/blog/face_recognition_with_opencv2/

 

 

第二部分:人脸识别最新趋势汇总

1、基于深度学习的人脸识别技术综述

简介:本文综述了8种基于深度学习的人脸识别方法,包括:1,face++(0.9950 );2,DeepFace(0.9735 );3,FR+FCN(0.9645 );4,DeepID(0.9745 );5,FaceNet(0.9963 );6, baidu的方法(0.9977 );7,pose+shape+expression augmentation(0.9807);8,CNN-3DMM estimation(0.9235 )。述方法可以分为两大类:第一类:face++,DeepFace,DeepID,FaceNet和baidu。他们方法的核心是搜集大数据,通过更多更全的数据集让模型学会去识别人脸的多样性。第二类:FR+FCN,pose+shape+expression augmentation和CNN-3DMM estimation。这类方法采用的是合成的思路,通过3D模型等合成不同类型的人脸,增加数据集。这类方法操作成本更低,更适合推广。

参考网址:https://zhuanlan.zhihu.com/p/24816781

 

2、基于mtcnn和facenet的实时人脸检测与识别系统开发

简介:本文主要介绍了实时人脸检测与识别系统的详细方法,该系统基于python/opencv2/tensorflow环境,实现了从摄像头读取视频,检测人脸,识别人脸的功能。实现类似功能的代码有openface,但是,openface核心是基于torch和lua。

参考网址:https://zhuanlan.zhihu.com/p/25025596

 

3、谷歌人脸识别系统FaceNet解析

简介:作者开发了一个新的人脸识别系统:FaceNet,可以直接将人脸图像映射到欧几里得空间,空间距离的长度代表了人脸图像的相似性。只要该映射空间生成,人脸识别,验证和聚类等任务就可以轻松完成。文章的方法是基于深度卷积神经网络。FaceNet在LFW数据集上,准确率为0.9963,在YouTube Faces DB数据集上,准确率为0.9512。

参考网址:https://zhuanlan.zhihu.com/p/24837264

 

第三部分:人脸识别图片库资源

1、Labeled Faces in the Wild

简介:A database of face photographs designed for studying the problem of unconstrained face recognition. The data set contains more than 13,000 images of faces collected from the web. Each face has been labeled with the name of the person pictured. 1680 of the people pictured have two or more distinct photos in the data set.

网址:http://vis-www.cs.umass.edu/lfw/

 

2、李子青团队的webface

简介:该数据集合需要申请,个人申请无效,需要学校部门的领导或者代表。其中包含10,575个人,494,414幅图像。

 

3、ImageNet

简介:ImageNet is an image database organized according to the WordNet hierarchy (currently only the nouns), in which each node of the hierarchy is depicted by hundreds and thousands of images. Currently we have an average of over five hundred images per node. We hope ImageNet will become a useful resource for researchers, educators, students and all of you who share our passion for pictures.

官网地址:http://www.image-net.org/

资讯:http://www.image-net.org/challenges/LSVRC/2012/

数据集下载(百度网盘):链接: https://pan.baidu.com/s/1pKXjaZp 密码: c7pr

 

第四部分:领域专家及企业

Face++:https://www.faceplusplus.com.cn/

商汤科技sensetime:http://www.sensetime.com/

知图:http://www.tusimple.com/?lang=en

Linkface:https://www.linkface.cn/index.html#cases

格灵深瞳:http://www.deepglint.com/

中科奥森( 李子青为创始人):http://www.authenmetric.com/

云从科技:http://www.cloudwalk.cn/api_0.html

腾讯优图:http://open.youtu.qq.com/

宇泛科技Uni-Ubi:http://www.uni-ubi.com/

彭春蕾:http://chunleipeng.com/

汤晓鸥教授+王晓刚教授:http://mmlab.ie.cuhk.edu.hk/

李子青老师:http://www.cbsr.ia.ac.cn/users/szli/

马毅老师:http://sist.shanghaitech.edu.cn/StaffDetail.asp?id=3

张磊:http://www4.comp.polyu.edu.hk/~cslzhang/

Face Recognition Vendor Test (FRVT)https://www.nist.gov/programs-projects/face-recognition-vendor-test-frvt

Deep Learning(An MIT Press book): https://www.deeplearningbook.org/

 

发表评论

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen: