M1-T Point Cloud Unsampling and Compression (Point Cloud Denoising)

PI: Zhu Li

The goal of this project is to advance the point-cloud post-processing using deep learning method to understand the global and local manifolds of a 3D object.

Methods

  • Use deep learning to learn the global and local manifolds of a 3D object.
  • Estimate local surface from the learnt features.
  • Point Cloud in-painting and super resolution.
  • Dynamic Point Cloud Compression.
  • Inter-frame Point Cloud Motion Estimation

Deliverables

  • Algorithm research and evidences in publications.