Umberto Castellani and Adrien Bartoli
This repository contains python3 code implementation of the 3D shape registration article by Dr. Castellani and Dr. Bartoli
Although the main focus of this literature is 3D registration, the concept can be extrapolated for 2D registration as well. As a matter of fact, 2D registration is used to check the correctness of this implementation. 2D registration can be visualized by running the file two_dimension_registration_demo.py.
cd demo && python3 two_dimension_registration_demo.py
Given two dataset, where one is the rigidly translated and rotated form of another, ICP can be used to find the translation and rotation matrices to fit one dataset on another. A demo can be run using
cd demo && python3 three_dimension_registration_demo.py
Stanford bunny model is an open source point cloud scan of a bunny statue. The point clouds are available in four resolutions containing 453, 1889, 8171 and 35947 points.
cd demo && python3 three_dimension_registration_demo.py --stanford-bunny
The point cloud with 1889 points are used as the input for this demo.