Point Cloud Central

A playground for experimenting with point cloud processing techniques, from feature extraction to surface reconstruction.

What’s Inside

A series of notebooks that build on each other:

• Geometry feature extraction — Computing local geometric properties from raw point data
• Optimal neighborhood selection — Finding the right scale for local analysis by minimizing eigenentropy
• Surface point extraction — Isolating surface points from volumetric data using normal-based detection
• Surface reconstruction — Turning scattered points back into continuous surfaces

Why Point Clouds

Point clouds are everywhere: LiDAR scans, depth cameras, medical imaging, and simulation outputs. But they’re just unstructured sets of coordinates. Extracting meaningful geometry from them requires careful processing, and this repo is where I test different approaches.