GilbertRC / ProLiF

[Code ongoing] Progressively-connected light field network for efficient view synthesis

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

ProLiF

We present a novel neural light field representation for efficient view synthesis.

This is the official repo for the implementation of Progressively-connected light field network for efficient view synthes.

How to run?

Setup

Clone this repository

git clone https://github.com/Totoro97/ProLiF.git
cd ProLiF
pip install -r requirements.txt

Training

  • Training for novel view synthesis
python exp_runner.py --config-name=prolif case_name=<case_name>
  • Training for scene fitting under varing light conditions
python exp_runner.py --config-name=prolif-lpips case_name=<case_name>
  • Training for text-guided scene style editing
python exp_runner.py --config-name=prolif_clip case_name=<case_name>

Testing

  • Rendering all test views
python exp_runner.py --config-name=<config_name> case_name=<case_name> mode=validate is_continue=true  # use latest checkpoint

The synthesized images can be found in exp/<case_name>/<exp_name>/validation.

  • Rendering video
python exp_runner.py --config-name=<config_name> case_name=<case_name> mode=video is_continue=true  # use latestcheck point

The synthesized video can be found in exp/<case_name>/<exp_name>/video.

Train ProLiF with your custom data

We follow the same data convention as LLFF. You may follow the original LLFF instruction for data preparation.

Citation

Cite as below if you find this repository is helpful to your project:

@article{wang2021prolif,
  author    = {Wang, Peng and Liu, Yuan and Lin, Guying and Gu, Jiatao and Liu, Lingjie and Komura, Taku and Wang, Wenping},
  title     = {Progressive-connected Light Field Network for Efficient View Synthesis},
  journal   = {Arxiv},
  year      = {2022},
}

About

[Code ongoing] Progressively-connected light field network for efficient view synthesis


Languages

Language:Python 100.0%