gilbaz / LORAX

CVPR Paper - "3D Point Cloud Registration for Localization using a Deep Neural Network Auto-Encoder" - Partial Implementation Code
91 stars 29 forks source link

LORAX

CVPR Oral Presentation Paper - "3D Point Cloud Registration for Localization using a Deep Neural Network Auto-Encoder"

Authors: Gil Elbaz, Tamar Avraham, Prof. Anath Fischer

Technion - Israel Institute of Technology

*An RSCS implementation is included. In addition a fixed-RSCS algorithm is also implemented. It is an augmentation of the original algorithm that allows for better parametric control. Using this you can define exactly how many Super-points should be created, how many points in each Super-point.

*The basis code for the Super-point Auto-encoder Feature generator will be included.

*Requires Python 3.5 or higher