DEV Community

Cover image for Using AI to avoid touching your face
All Around Azure

Using AI to avoid touching your face

Frank Boucher ☁
Cloud AdvocateπŸ₯‘ passionate about the cloud, supporting developer communities with #Azure 🎦 in #FR & #EN
Originally published at ・2 min read

πŸ“Ί 028 - Em Lazer Walker πŸŽ™οΈ Frank Boucher

In this episode, Em is showing one of her last projects. She trained a machine learning model using Azure Custom Vision to detect when someone, facing the camera, is touching or not their face. All this project is also available in GitHub.

Em will starts by explaining what a good machine learning model, and explain why the one use, in this case, is only "good enough for a demo". We will cover different options available when building your model and the scenarios when those options should be used. Once created, we will train our brand new model using many images of faces for both of the scenarios: touching and not touching.

The second part of this demo is composed of a JavaScript webapp. We will walk through the code that does all the operations to grab an image from the video feed and send it to the Custom Vision service.

Finally, she will share a VR project, inspired by the Microsoft Flight simulator (coming out that day), that use the Bing map service to let us fly in any sky around the globe.

Useful Links

πŸ”— Custom Vision

πŸ”— Custom Vision Quickstart

πŸ”— How YOU Can Use Computer Vision to Avoid Touching Your Face!

πŸ”— Don't Touch Your Face app

πŸ”— Don't Touch Your Face code

πŸ”— Bing Map Service

Learn more about Em Lazer Walker

πŸ”— @lazerwalker

πŸ”— Bio

Find us on

πŸ“Ί Twitch

πŸ”— Twitter


Discussion (0)

Forem Open with the Forem app