Real time face recognition with CPU
source link: https://www.tuicool.com/articles/y6bUzuN
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
After reading this article, you will understand how to use pre-trained models to build a real-time face recognition system with CPU.
Oct 31 ·6min read
Both the academic and industrial fields are putting in tremendous efforts to develop face recognition algorithms and models that are both, fast and accurate. Thanks to these efforts, it is now possible to accomplish accurate, real-time face recognition for multiple faces with CPU.
In this article, we will compose a real-time face recognition system with the Ultra-light face detector by Linzaer and MobileFaceNet¹.
Face Detection
In order to recognize a face, we would first need to detect a face from an image. There are many ways to do so.
I have explored multiple face detectors. These include Face-recognition packge (containing Histogram of Oriented Gradients (HOG) and Convolutional Neural Network (CNN) detectors), MTCNN , Yoloface ,Faced, and a ultra light face detector released recently.
I found that while Yoloface has the highest accuracy and most consistent execution time, the Ultra-light face detector was unrivalled in terms of speed and produces a relatively good accuracy.
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK