We just published a deep technical piece on Medium - about the biggest hidden limitation of standard AR.

There's a popular technology called Augmented Image: you anchor a 3D model to a real-world image, and it "comes to life." Works out of the box on Android and iOS. Great for gallery paintings, posters, product packaging.

But it has two problems:

1. Only works within 5 meters. So a painting on a wall - yes. A logo on a skyscraper - no.

2. Doesn't recognize "simple" images. Minimalist logos, monochrome compositions - the standard algorithm just misses them.

And what do we work with? AR installations on real buildings in Dubai, including the Burj Khalifa. The standard stack doesn't cut it here.

We spent six months figuring out how to get around this.

In the article:

▫️ Why ARCore and ARKit work the way they do
▫️ What we tried with neural networks
▫️ Two breakthroughs we've achieved
▫️ Two problems we haven't solved yet
▫️ Where all of this is heading

Full article read 👉 here

If you're in AR - drop a comment, let's talk.

AR Spatially