Ubuntu20.04复现MVSDF(ICCV2021):学习多视图表面重建的有符号距离场

目录

参考

一、环境配置

二、数据集下载

三、训练(以DTU_scan24为例)

四、测试(以DTU_scan24为例)


参考

论文链接:https://arxiv.org/pdf/2108.09964.pdf

项目代码链接:https://github.com/jzhangbs/MVSDF?tab=readme-ov-file

一、环境配置

dependencies:
  - cudatoolkit=10.2.89
  - numpy=1.19.2
  - python=3.8.8
  - pytorch=1.7.1
  - tqdm=4.60.0
  - pip:
    - cvxpy==1.1.12
    - gputil==1.4.0
    - imageio==2.9.0
    - open3d==0.13.0
    - opencv-python==4.5.1.48
    - pyhocon==0.3.57
    - scikit-image==0.18.3
    - scikit-learn==0.24.2
    - trimesh==3.9.13
    - pybind11==2.9.0

 其中pytorch安装参考:

pip install torch==1.8.0+cu111 -f https://download.pytorch.org/whl/cu111/torch_stable.html

二、数据集下载

下载预处理的DTU数据集:

https://hkustconnect-my.sharepoint.com/personal/jzhangbs_connect_ust_hk/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Fjzhangbs%5Fconnect%5Fust%5Fhk%2FDocuments%2Frelease%2Fmvsdf%5Fdata%2Ezip&parent=%2Fpersonal%2Fjzhangbs%5Fconnect%5Fust%5Fhk%2FDocuments%2Frelease&ga=1

数据集文件夹内容:

MVSDF(总文件夹名称)
 |
 | code
 |     | confs
 |     |  ...
 |     |  ...
 |     | mvsdf_data(数据集文件夹)
 |     |  |-- scan24
 |     |         |-- cam_00000000_flow3.txt      
 |     |         |--        ...
 |     |         |-- imfunc4  
 |     |         |-- pair.txt
 |     |  |-- scan37
 |     |      ...
 |      
 | media
 | ...
 | ...

三、训练(以DTU_scan24为例)

训练指令:

cd code   #进入含有代码的文件夹
python training/exp_runner.py --data_dir mvsdf_data/scan24/imfunc4 --batch_size 8 --nepoch 1800 --expname dtu_scan24

 训练结果:

四、测试(以DTU_scan24为例)

测试指令:

python evaluation/eval.py --data_dir mvsdf_data/scan24/imfunc4 --expname dtu_scan24 --eval_rendering

测试结果

猜你喜欢

转载自blog.csdn.net/weixin_59961223/article/details/135439416