Tag: 3d
All the articles with the tag "3d".
Semantic Segmentation of Point Clouds with PointNet++
Published: at 12:30 PMA practical guide to semantic segmentation of 3D point clouds with PointNet++. Learn why point clouds are hard for neural networks, how PointNet++ builds a hierarchical feature representation with sampling and grouping, and how to implement the key building blocks in PyTorch, with a worked example on the S3DIS indoor dataset.
Gaussian Splatting vs. Traditional Photogrammetry
Published: at 10:00 AMA technical comparison between 3D Gaussian Splatting and traditional photogrammetry. Both techniques reconstruct 3D scenes from photographs, but differ fundamentally in their approach, strengths, and ideal use cases.
Point Cloud Registration with Open3D: A Complete Guide
Published: at 10:00 AMA comprehensive guide to point cloud registration using Open3D in Python. Learn about ICP (Iterative Closest Point), global registration with RANSAC and FPFH features, and how to build a complete registration pipeline for 3D data alignment.