Imagine seeing a tiny sphere behind a foggy pane and, with the help of AI, pinpointing its location to the theoretical limits of physics. That’s exactly what scientists from TU Wien and partners have achieved.
Key Points at a Glance
- AI-trained neural networks can infer position of hidden objects hidden behind turbid media nearly at Fisher information limits.
- The team established theoretical boundaries for optical measurement precision, then built an experiment to test them.
- Algorithmic approach opens possibilities for enhanced medical imaging, materials research, and quantum technologies.
For over 150 years scientists have known there’s a hard cap on how sharp optical imaging can be—nature itself enforces a blur. At TU Wien, along with the University of Glasgow and Grenoble, researchers decided to ask: can modern AI actually reach that cap?
They began by formally quantifying the theoretical limit of precision using Fisher information—a concept telling how much information your sensor has about a hidden parameter, like an object’s position. If that metric is low, no amount of clever image processing can help. This sets the “Cramér–Rao bound”—the absolute threshold of achievable accuracy.
Next, experimentalists built a clever setup: a laser illuminates a tiny reflective sphere hidden behind turbid glass or liquid. On the opposite side, a camera captures wildly distorted, seemingly random light patterns. These diffraction patterns show no obvious clues to the human eye on where the object actually sits.
Enter neural networks. By feeding thousands of these distorted images—each labeled with the known position of the sphere—the AI begins to connect light texture with spatial displacement. It learns the hidden code embedded within the blur.
Once trained, the AI delivered results that were remarkable: its position estimates were only marginally worse than the theoretical limit derived via Fisher information. In other words, it decoded enjoyment from noise almost as optimally as physics allows.
“Our AI-supported algorithm is not only effective, but almost optimal,” says Prof. Stefan Rotter from TU Wien. “It achieves almost exactly the precision that is permitted by the laws of physics.” The few-percent discrepancy marks state-of-the-art performance in optical metrology.
Why does it matter? Imagine medical imaging—from deep-tissue scans to endoscopy—where light scattering blur has long masked fine detail. Or materials science, where internal flaws or stress points are hidden deep within. Or even quantum technologies where tiny positional shifts matter. This tool could bridge those precision gaps.
Next steps? The researchers plan to collaborate with medical and applied physics partners to test this AI model in real-world imaging contexts—from bio‑tissue diagnostics to micro‑imaging in materials labs.
Source: TU Wien News
Enjoying our articles?
We don’t show ads — so you can focus entirely on the story, without pop-ups or distractions. We don’t do sponsored content either, because we want to stay objective and only write about what truly fascinates us. If you’d like to help us keep going — buy us a coffee. It’s a small gesture that means a lot. Click here – Thank You!
