Lingjing324 / PR-Net

Non-Rigid Point Set Registration Networks
https://arxiv.org/abs/1904.01428
MIT License
53 stars 13 forks source link

PR-Net

(1) The code is tested on: python 3.6

pytorch: conda install pytorch torchvision cudatoolkit=8.0 -c pytorch

(2) For training the model: python train.py

(3) C.D. (Before and After) is saved in Vis_xxx and a sample of qualitative testing registration results is saved in Vis_xxx.mat.

(4) For comparison with CPD, we use the package: https://github.com/siavashk/pycpd

(5) geotnf is modified based on the code from paper: Convolutional neural network architecture for geometric matching,Ignacio Rocco, Relja Arandjelović, Josef Sivic.