NVIDIA Neuralangelo, from video to 3D models

we can evaluate artificial intelligence like NVIDIA Neuralangelo as a “renaissance”? The truth is that both his title and the first of the visual examples they chose to illustrate the potential of this new development point in this sense, not so much in the sense of suggesting that he has skills in all sorts of fields, something that Leonardo da Vinci the biggest exemplar, such as that which refers to the impressive works that geniuses like Michelangelo were able to create with a chisel in hand from a large block of marble, to whom the name of this new AI model is of course dedicated.

And what made Neuralangelo deserve such a description? Well, based on conventional video footage (that is, in 2D) that surrounds an object of any type or that passes through a room, borders a building, etc. will be able to create a high-quality three-dimensional representation of itwhich, in addition to the shape of the surfaces, also takes into account the type of material used, which will be extremely practical.

Once the model is generated, the technology developed by NVIDIA Research allows it to be exported, so in this way the model can be commonly used in 3D design applications, either because we want to include an object in a previously designed scene, or because we want to create a scene from this model. According to Ming-Yu Liu, co-author of the Neuralangelo paper, eventually it will be possible to import models with any level of detailwhich prompts us to think that the texture information captured and recreated in the process will be preserved.

To create a 3D representation, NVIDIA Neuralangelo analyzes the video footage once it locates what you want to capture andchoose a selection of boxes that provide all the information you need in terms of depth, size and shape. Next, it will proceed to create a first approximate three-dimensional representation, which will be optimized in the third stage to faithfully reproduce all the details of the object it captures.

This is not the first time NVIDIA has explored an AI algorithm of this type. Actually, some of the technology from the Instant NeRF model we told you about early last year was used to develop Neuralangeloalthough this latest development provides a more detailed and realistic result, we can understand that both thanks to the improvements in the algorithm itself and the fact that it will have more information about the object to be reproduced than that provided by the images used through Instant NeRF.

More information