Algorithms and data structures for AI and simulation applications on sparse 3D data.
We have developed and open-sourced several SDKs for efficient storage and computation on large sparse 3D data like meshes, point clouds, Gaussian splats, implicit surfaces, CAD geometry, and more. These libraries support GPU-accelerated applications like deep learning surface reconstruction, segmentation, training and rendering of neural radiance fields, physics simulations, level set methods, and much more.
Spatial Intelligence • Sparse Volumes • Simulations • Neural Compression
fVDB is an open-source deep learning framework for sparse, large-scale, high-performance spatial intelligence. It builds NVIDIA-accelerated AI operators on top of NanoVDB to enable reality-scale digital twins, neural radiance fields, 3D generative AI, and more.
NanoVDB is an open-sourced data structure and toolset for efficient representation, rendering, and simulation of sparse volumetric 3D data. It is GPU-accelerated and part of the OpenVDB project that has become an industry standard for sparse volumes.
NeuralVDB is a large-scale volume representation with AI-enabled data compression technology. It introduces lossy compression to OpenVDB, the industry-standard library for simulating and rendering sparse volumetric data such as water, fire, smoke and clouds.