NVIDIA has introduced a new AI model called Neuralangelo that can create 3D replicas of objects from 2D videos, whether they’re classic sculptures or run-of-the-mill trucks and buildings. Neuralangelo works by choosing a few frames that show a subject in a 2D video from different angles, so you can get a clear idea of the subject’s depth, size and shape. Next, create a rough 3D representation of your object, then optimize it to mimic real-world details.
According to the company, the new model adopts the technology of the old model “Instant NeRF”, which can accurately capture the details of what users want to recreate in 3D. These include variations in textures, patterns and colors. NVIDIA says Neuralangelo’s ability to capture tricky textures, such as the roughness of roof shingles and the smoothness of marble, “is significantly superior to conventional methods.” It also seems to have enough capabilities to create usable virtual objects from videos taken with smartphones, so creators could easily generate 3D recreations of him for their projects. This is also a quick solution. According to NVIDIA Research, the tool will allow him to create his 3D scene in less than two hours.
Ming-Yu Liu, co-author of the paper on Neuralangelo, said, “Eventually, developers will be able to transform detailed objects, whether small statues or gigantic buildings, into virtual environments for video games and industrial digital twins. can be imported into . In the video below showing what the technology can do, NVIDIA says it can create large-scale scenery from drone footage, making it easier for developers to recreate real-world locations in-game. . There are many other potential uses, such as robotics, virtual reality experiences, architecture and, of course, art.
The company’s research division plans to unveil Neuralangelo at its computer vision and pattern recognition event June 18-22 in Vancouver.
