AI software creates 3D models from videos

Jaw setknown for efficient voxel-based 3D modeling tools, has introduced a new AI-powered 3D scanning tool called Postshot in beta. This software solution uses AI techniques to generate detailed 3D models from video recordings.

Unlike hardware-based 3D scanners that use special sensor systems, Postshot relies on smartphone cameras and complex software algorithms. The user simply records a video of the object to be scanned and imports it into the Postshot desktop application.

Postshot combines two modern AI technologies: NeRF (Neural Radiance Fields) and Gaussian Splatting. NeRF trains an AI model on the object’s radiance and can use it to generate views from all angles. Gaussian Splatting, on the other hand, generated solid models without the use of surfaces or lines.

The software offers different approaches to convert the video into a 3D model. This flexibility enables a wide range of applications in different areas of 3D modeling and visualization.

Postshot is currently free to use as it is a beta version. However, users need a Windows platform with a powerful GPU.

Although not specifically designed for 3D printing, Postshot shows the potential of new AI methods for creating 3D models. Compared to traditional photogrammetry, these techniques can fill gaps more intelligently and produce detailed models.