#This repository contains my experience of using it

#If you have the computing power feel free to make your own model or use a pre-trained one over a huge dataset

#Credits go to Mr Nathan Bond who streamlined the process of learning for me. I am extremely grateful to him for his time Here is a an article made by him which will make it easy for users. https://medium.com/@nhbond/how-to-copying-what-david-sandberg-did-4bf6e0ad3cd5

#Observations: 1)It correctly identifies images even if test images are placed in the wrong folder. 2)It's accuracy increases if trained over more images of a single person. 3)If a set of white images or unrelated images are put, there is still detection but accuracy is very low. This proves that there must be a minimum threshold. 4)Actual math algorithms are executed in high performing C++. Will learn about them in the next few weeks. 5)Excited to see how python creates a flow for the math to be executed.

#Links 1)Original by Mr Sandberg - https://github.com/davidsandberg/facenet 2)Link to pre-trained model - https://drive.google.com/file/d/1EXPBSXwTaqrSC0OhUdXNmKSh9qJUQ55-/view 3)Link to addtional reading and alternative methods - https://github.com/davidsandberg/facenet/wiki