iPhone X Face ID hits people with surprise, people starting to realize through AI Deep learning their face is much more unique and accurate than their fingerprint. But what some people don't realize is the reason why iPhone X facial recognition works is that it's only detecting you or not you, therefore it is at much higher accuracy than using AI to detect multiple target.

What it does

We've built a platform that shows how the AI works on Intel Movidius NCS, using all the default camera that came with the development kit. This project can be extended to using facial recognition to unlock deadbolts, record entries, turn on different light themes, and many others. Working with Walabot Radar, we can have additional securities for those tries to fool the camera.

How I built it

We've trained the entire network on caffe, reaching more than 99% accuracy with "me or not me" method. Additional radar (Walabot) will be added to the project to make sure that simple images can not pass the test. I wrote a 9 step guide which can be seen at link

Challenges I ran into

Getting entire product to work on IoT was rather difficult

Accomplishments that I'm proud of

Writing up 2000 word guide on how to build the entire guide, this can be seen at link

What I learned

How to integrate AI with IoT and Alexa seamlessly for the user

What's next for AI Facial Recognition Lock

Improving the algorithm, if we get enough interest we can build this into an actual product. Our Alexa skill is published as Face Lock under skill id for those who are interested in taking a look

Built With

  • alexa
  • lambda
  • movidius-ncs
  • up2
  • walabot
Share this project: