Sharing of four open source face recognition projects

Face recognition is a technology that can recognize or verify people in real time from the video frames of images or video sources . This article shares four open source face recognition projects , To improve your skills in Data Science .

Be careful ： This article is just a brief introduction to some less famous but very good open source projects , You can use these projects in your projects .

1. Face recognition

Adam Geitgey Of Face_Recognition It is the simplest built-in in the world Python Face recognition API, You can use... From the command line . This project is based on deep learning , Use dlib The most advanced face recognition library .

It's called “ The most simple ”, Because it allows you to add images to folders , And recognize faces from the command line , stay wild Mark the face in the benchmark , Accuracy rate is 98.38%

Face recognition API Other features of include ：

• Detect multiple faces in a picture , And identify the people in each photo .
• Detect faces in real-time cameras .
• Detect human eyes 、 nose 、 mouth 、 eyebrow 、 Facial features such as chin , Get the position and contour of the detected part .
• Detect facial features and apply digital makeup

2. DocFace

DocFace Is an open source face recognition system , It can be used to match ID photos with self taken photos in real time . This project is based on TensorFlow and Python Above .

To ensure better performance , use first MatLab Version of MTCNN following SphereFace Align the facial selfie taken , The data set used to train the basic model is Ms-Celeb-1M and LFW.

Then use the basic model to learn about ID Fine tune the self timer dataset . Learning through transfer , Use the pre trained basic model , We can achieve 99.67% The accuracy of .

3. GetMeThrough

GetMeThrough It's a free open source software , Working in real time in offline mode web Applications , Help the organizer of any event to allow only authorized or invited people to participate in the event , Use a two-step validation factor , That is, first use face recognition technology to check whether the person is registered in the database , Otherwise, the QR code will be checked .

This project uses dlib Pre training model construction , The model is based on Face_Recogniton API（ As mentioned earlier ） above , In order to achieve 99.38% The accuracy of . Other tools used in the development of this project include MongoDB、materialecss,Node.js as well as Express.js For the front end 、 Back end 、 Database and web Application framework .

Follow the instructions given here , You can get a copy of the project running on your local computer , For development and testing .

4. SharpAI DeepCamera

sharpAI Of DeepCamera yes Android Open source AI video surveillance on devices , The surveillance camera has face recognition 、 Human shape recognition 、 Motion detection 、 Face detection 、 Target detection and other functions .

This is a free automatic machine learning （AutoML） Edge AI platform for deep learning , On this platform , Training a new model does not require programming experience , It is mainly used to protect your privacy .

It has been supported in various Android The device and camera work well . at present ,DeepCamera from SharpAI maintain .

Link reference

Face_Recognition API

• https://github.com/ageitgey/face_recognition
• https://medium.com/@ageitgey/machine-learning-is-fun-part-4-modern-face-recognition-with-deep-learning-c3cffc121d78

DocFace

• https://arxiv.org/abs/1805.02283
• https://github.com/seasonSH/DocFace

Get Me Through

• https://github.com/malikshubham827/get-me-through

DeepCamera

• https://github.com/SharpAI/DeepCamera
• https://sharpai.github.io/DeepCamera/