Design researcher Anuradha Reddy has come up with a unique take on physical security keys, with the CryptoCrochet-Key — a crocheted object which pairs with a computer vision system built on Teachable Machine to offer a unique authentication device for the smart home.
"Make your own uniquely unique keys to secure your home IoT AI devices using Crypto 8-ply yarn and machine learning image classification with Teachable Machine," Reddy writes of the work-in-progress project. "The ‘crypto’ yarn is the name given to this colorful yarn which inspired the project and has little to do with actual cryptography. Still, anything you make from this yarn will be a unique object, which can in theory be useful from security POV."
The core concept is deceptively simple: A crochet pattern which, when followed, results in a soft object the shape of a key. While each key made using the pattern will be almost exactly the same shape, the Crypto yarn means it will have different colors in different sections of the key — and, in theory, no two keys should ever be alike.
Once a CryptoCrochet-Key has been made, it can be used as a unique identifier through any webcam and a machine learning system built on Google's Teachable Machine platform. "[I] have plans to port this ML model to a [Raspberry Pi] with a camera," Reddy adds, "so stay tuned!"
Reddy has published the crochet pattern on Twitter, with more details available in the project thread; the Teachable Machine has not yet been released, but will follow for what Reddy describes as "an easy project [for] a weekend."