Skip to content

Latest commit

 

History

History
53 lines (39 loc) · 1.43 KB

README.md

File metadata and controls

53 lines (39 loc) · 1.43 KB
Raw

PCLossNet

[ECCV'22] The codes for Learning to Train a Point Cloud Reconstruction Network without Matching

Environment

  • TensorFlow 1.13.1
  • Cuda 10.0
  • Python 3.6.9
  • numpy 1.14.5

Dataset

The adopted ShapeNet Part dataset is adopted following FoldingNet, while the ModelNet10 and ModelNet40 datasets follow PointNet. Other datasets can also be used. Just revise the path by the (--filepath) parameter when training or evaluating the networks. The files in (--filepath) should be organized as

    <filepath>
    ├── <trainfile1>.h5 
    ├── <trainfile2>.h5
    ├── ...
    ├── train_files.txt
    └── test_files.txt

where the contents in (train_files.txt) or (test_files.txt) should include the directory of training or testing h5 files, such as:

    train_files.txt
    ├── <trainfile1>.h5
    ├── <trainfile2>.h5
    ├── ...

Usage

  1. Preparation
cd ./tf_ops
bash compile.sh
  1. Train

For the reconstruction task,

Python3 vv_ae.py

Note that the paths of data should be edited through the (--filepath) parameter according to your setting.

  1. Test

For the evaluation of reconstruction errors,

Python3 vvae_eva.py

The trained weight files should be provided by the (--savepath) parameter to evaluate the performances.