Skip to content

dude123studios/SmarterRingV2

Repository files navigation

SmarterRingV2

Project Image

This is an aplied deep learning project that will literally tell you who is at your door, once the doorbell is rung.


Table of Contents

Medium Blog: https://dude123studios.medium.com/face-recognition-at-your-doorbell-1e6f250f2ee9


Description

Face Recognition is huge. With the ability to automatically classify someone, many useful prducts have been created. For example, to unlock the new IPhones all you need is to display your face. This single handedly shows how strong our technology has come. Millions rely on a deep clustering model to protect all of their private information and other smart phone data. With the new Amazon Ring, I was disapointed that it didn't have this common feature. So I decided to create it myself! Read my medium blog on the subject. This repository can be used on your own Amazon Ring in just a couple of steps! Diagram


Technologies

  • Deep Learning
  • Ring API
  • Computer Vision
  • Face Recognition

If you are interested in FaceNet the model used in this repository, the following diagram illustrates it: Architecture

Back To The Top


Installation

  $git clone https://github.com/dude123studios/SmarterRingV2
  $pip install requirements.txt
How To Use

Prerequisites Set environment variables, USERNAME and PASSWORD, and equate them to your Amazon Ring acount's username and password

Gather Data In order to submit the faces of your family members, or those who you would like to have regonized, you have two options. (name = name of person to enter)

  • run:
  $python submit_face.py name False
  • or:
  $python submit_face.py name True

In the first case, a popup will show on your screen using your webcam. Hit the space bar when ready to record an image of "name". Do so multiple times from different angles. Once done, hit the escape button to close the program. In the second case, data will be grabbed directly from your ring doorbell."name" is the name of the person who was at your doorbell last when it was rung. This method will reach higher accuracy.

Encode Data
Once all the images are in, run:

  $python create_encodings.py

This will simply encode your data.
Run
Finally, to run the program, run main.py. Leave it running. Once anyone rings your doorbell, you will have a loud Text-To-Speech played automatically on your device informing you who it is.

  $python main.py

Test(Optional)
run:

  $python test.py

A window will pop up showing your webcam detecting your face. Use incase you run into any issues or want to test the capabilities of the program

Back To The Top


References


Author Info

  • Name: Atharv N.
  • Location: California USA
  • Age: 13
  • Hobbies: Math, Programming, Minecraft
  • Email: dude123studios@gmail.com

Back To The Top

Releases

No releases published

Packages

No packages published

Languages