Researchers from the University of Maryland have developed a method that uses a NeRF AI neural network capable of reconstructing 3D images from eye reflections. This is possible by using a high-resolution camera to capture multiple images in a fixed position, focused on an individual in motion looking towards the camera.
When zooming in on this passport or driver’s license-sized image, a reflection should be visible in the person’s eye, with a mirror image as well as objects in the area identifiable. Removing the iris from the reflection required=s texture decomposition by the AI, which meant it had to be trained on a 2D texture map that learns its surface before it can remove it.
- Transform your reality with Meta Quest 3S 128GB. Now get the Amazon-exclusive Cardboard Hero Bundle, which includes the Handiwork Helmet, Handiwork...
- Turn any room into your own personal theater. Dim the space around you and watch on a giant, vibrant screen. Go all in with USB-C headphones, or plug...
- Have more fun with friends in Quest. Whether you’re stepping into an immersive game with people from around the world, watching a live concert...
We further propose a simple regularization prior on the iris texture pattern to improve reconstruction quality. Through various experiments on synthetic and real-world captures featuring people with varied eye colors, we demonstrate the feasibility of our approach to recover 3D scenes using eye reflections,” said the researchers.
