Title: Mapping 3D Shapes To A Human-Interpretable Object

About This Project:

GeoCode maps 3D shapes to a human-interpretable parameter space, allowing to intuitively edit the recovered 3D shapes from a point cloud or sketch input. The teams has used a neural network to map a given point cloud or sketch to our interpretable parameter space. Once produced by procedural program, shapes can be easily modified. Empirically, the team shows that GeoCode can infer and recover 3D shapes more accurately compared to existing techniques and we demonstrate its ability to perform controlled local and global shape manipulations.

Requirements:

• Python 3.8

• CUDA 11.8

• GPU, minimum 8 GB ram

• During training, a machine with 5 CPUs is recommended

• During visualization and sketch generation, we recommend a setup with multiple GPU nodes, refer to the additional information to run in parallel on all available nodes

• During test-set evaluation, generation of raw shapes for a new dataset, and during stability metric evaluation, a single node with 20 CPUs is recommended

Project Demo Video:

Code And Other Details: Mapping 3D Shapes To A Human-Interpretable Object

Related Projects:

• Build A Self-Driving Car Like NVIDIA

• Text To 3D Stable Dream Fusion

• Build A 3D Data Visualizer