What's going on?

Using Tensorflow.js, we're able to use deep learning to detect objects from your webcam! Your webcam feed never leaves your computer and all the processing is being done locally! (Trust me, we can't afford the servers to store/process your data)

Can I use something like this in my project?

Yes! Check out it out on ModelDepot!

Why is it so slow?

Real-time object detection is a challenging task, and most models are optimized to run fast on powerful GPU-powered computers with optimized code. Most of us don't have super fast GPUs (especially if you're browsing on mobile) and Tensorflow.js can't take full advantage of our computer's GPUs.

How does this work?

We're using Tiny YOLO, a ML model that's been converted to work in Tensorflow.js.