There’s a saying that the eyes are the window to a person’s soul, but eyes may be a mirror, according to a team of researchers at the University of Maryland, who gathered enough data to recreate what a person was like in a short video clip. provides. Viewing as an interactive 3D model.
Spoilers of the Week | June 10
many factors CSI, and its spin-off, a popular franchise for CBS, but one that drew many viewers back week after week, was the top-notch technology crime scene investigators possessed. A grainy, low-resolution frame of video captured by a security camera in the middle of the night can be magnified repeatedly until it reveals individual fibers on a suspect’s shirt. Most of the tools used to solve crimes on the show still only exist in the writers’ imaginations, but researchers are making impressive strides when it comes to the amount of data that can be extracted from just a few frames of video footage.
Taking advantage of previous research on Neural Radiance Field (NERF) technology, where complex scenes or objects can be completely recreated in 3D using a partial set of 2D images captured at different angles, and the fact that Researchers at the University of Maryland were able to generate 3D recreations of simple scenes based on imagery extracted from reflections of the eye, despite the fact that the shape of the cornea of the eye is more-or-less the same for all healthy adults. However, don’t expect the results to be used to solve a crime.
This approach, as detailed in a recently published study, comes with some very unique challenges. For starters, the reconstruction of 3D models from 2D images usually starts with high-quality source material, such as video captured with a modern smartphone’s digital camera. Here, images are extracted from a small, low-resolution portion of each frame, and they are layered over the complex texture of the eye’s iris, whose color also varies from person to person, to be clearly defined. Requires extensive post-processing. Imagination.
Further complicating this approach is the fact that the series of 2D images being used for the reconstruction are all originating from the same location, with only minor changes being introduced as the subject’s eye looks around. . If you’ve ever tried to create a 3D model of an object or room using a smartphone camera and an app, you know that to get optimal results you have to move around and record it from all sides and angles Will happen. But that is not an option here.
As a result, 3D models created using this new technology are very low resolution and with little detail. As the researchers did, you could still identify objects like plush dogs, or bright pink Kirby toys, but these results were also achieved using optimal conditions: very basic scenes, very deliberate lighting, and High-resolution source imagery.
When the researchers tried to apply their approach to footage they didn’t capture, it was still created using ideal lighting conditions, such as a clip of Miley Cyrus. wrecking ball Music video sourced from YouTube, it is impossible to recognize what you are seeing in the resulting 3D model. This nebulous blob is likely a hole in the white shroud that the camera lens was looking through to achieve the distinctive look of this shot, but even Grissom’s CSI team could not confirm this using this 3D model. As fascinating as this research may be, it will be some time before it becomes a tool with practical application.
