I'm researching surface reconstruction from point clouds using machine learning. I have some ideas for improvement for my current work. It's currently in double-blind submission. Therefore, I can't tell too much here.
- Better architecture for the network
- Iterative evaluation for better performance and to avoid error propagation
- Regularization to get a smoother surface and get rid of noise
- Some low-level optimization
- Knowledge of English language (source code comments and final report should be in English)
- Basic knowledge of geometry for computer graphics (e.g. surface definition with vertices and faces, signed distance fields)
- Basic knowledge of Python
- Basic knowledge of Deep Learning (Pytorch)
- Knowledge of modeling and geometry processing
The source code is currently pure Python for both Windows and Linux.