This repository contains a custom implementation of PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space, originally proposed by Charles R. Qi et al: GitHub. The project is developed for educational purposes and as a personal practice project to explore PointNet++ for 3D point cloud segmentation/classification.
Status: The model is fully defined. The project is work in progress.
- ✅ Model architecture implemented end-to-end
- 🔧 Add dataloaders for multiple file types (e.g.,
.ply,.bin,.npz, etc.) - 🔧 Finalize
train.py(training loop, logging, checkpoints, eval) - 📚 Complete documentation and usage examples
- This code is intended for learning and experimentation; it is not production-ready yet.
- Interfaces and filenames may change during refactoring.
Qi, C. R., Yi, L., Su, H., & Guibas, L. J. (2017). PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space. NeurIPS 2017.
https://arxiv.org/abs/1706.02413
All credit for the original method to Charles R. Qi and co-authors. This repository aims to re-implement the core ideas for learning and practice.