Skip to content

hzxie/Pix2Vox

Repository files navigation

Pix2Vox

codebeat badge

This repository contains the source code for the paper Pix2Vox: Context-aware 3D Reconstruction from Single and Multi-view Images. The follow-up work Pix2Vox++: Multi-scale Context-aware 3D Object Reconstruction from Single and Multiple Images has been published in International Journal of Computer Vision (IJCV).

Overview

Cite this work

@inproceedings{xie2019pix2vox,
  title={Pix2Vox: Context-aware 3D Reconstruction from Single and Multi-view Images},
  author={Xie, Haozhe and 
          Yao, Hongxun and 
          Sun, Xiaoshuai and 
          Zhou, Shangchen and 
          Zhang, Shengping},
  booktitle={ICCV},
  year={2019}
}

Datasets

We use the ShapeNet and Pix3D datasets in our experiments, which are available below:

Pretrained Models

The pretrained models on ShapeNet are available as follows:

Prerequisites

Clone the Code Repository

git clone https://github.com/hzxie/Pix2Vox.git

Install Python Denpendencies

cd Pix2Vox
pip install -r requirements.txt

Update Settings in config.py

You need to update the file path of the datasets:

__C.DATASETS.SHAPENET.RENDERING_PATH        = '/path/to/Datasets/ShapeNet/ShapeNetRendering/%s/%s/rendering/%02d.png'
__C.DATASETS.SHAPENET.VOXEL_PATH            = '/path/to/Datasets/ShapeNet/ShapeNetVox32/%s/%s/model.binvox'
__C.DATASETS.PASCAL3D.ANNOTATION_PATH       = '/path/to/Datasets/PASCAL3D/Annotations/%s_imagenet/%s.mat'
__C.DATASETS.PASCAL3D.RENDERING_PATH        = '/path/to/Datasets/PASCAL3D/Images/%s_imagenet/%s.JPEG'
__C.DATASETS.PASCAL3D.VOXEL_PATH            = '/path/to/Datasets/PASCAL3D/CAD/%s/%02d.binvox'
__C.DATASETS.PIX3D.ANNOTATION_PATH          = '/path/to/Datasets/Pix3D/pix3d.json'
__C.DATASETS.PIX3D.RENDERING_PATH           = '/path/to/Datasets/Pix3D/img/%s/%s.%s'
__C.DATASETS.PIX3D.VOXEL_PATH               = '/path/to/Datasets/Pix3D/model/%s/%s/%s.binvox'

Get Started

To train Pix2Vox, you can simply use the following command:

python3 runner.py

To test Pix2Vox, you can use the following command:

python3 runner.py --test --weights=/path/to/pretrained/model.pth

If you want to train/test Pix2Vox-F, you need to checkout to Pix2Vox-F branch first.

git checkout -b Pix2Vox-F origin/Pix2Vox-F

License

This project is open sourced under MIT license.

About

The official implementation of "Pix2Vox: Context-aware 3D Reconstruction from Single and Multi-view Images". (Xie et al., ICCV 2019)

Topics

Resources

License

Stars

Watchers

Forks

Languages