Skip to content
No description, website, or topics provided.
Python Other
  1. Python 98.5%
  2. Other 1.5%
Branch: master
Clone or download
Latest commit 89565e3 Feb 12, 2020
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
siga19_source fixnet Feb 12, 2020
utils
COPYING
README.md update readme Feb 11, 2020
jointcap.bib add bib Feb 11, 2020

README.md

Thinking Lightstage

About

This is an implementation of the neural network proposed in Learning Efficient Illumination Multiplexing for Joint Capture of Reflectance and Shape.

Usage

System Requirements

  • Windows or Linux(The codes are validated on Win10, Ubuntu 18.04 and Ubuntu 16.04)
  • Python>=3.6
  • tensorflow>=1.11.0
  • numpy
  • opencv-python

Note: Since the training data (sampled points with BRDF parameters and position) are generated by an exe module, you may need a windows platform to generate training data.

Training

Windows:

  1. move to siga19_source
  2. run train.bat

Linux:

  1. move to siga19_source
  2. change DATA_ROOT in train.sh to the path of training data.
  3. run train.sh

Training Log & Model

When training is started, a log folder will be generated. You may run tensorboard to check training logs.
The test sample will be shown in the images tab in the order of:
input lumitexel, predicted diffuse lumitexel, ground truth diffuse lumitexel, ground truth specular lumitexel, predicted specular lumitexel
Trained lighting pattern can also be found in the images tab.
Trained model will be save in the created save folder.

License

Our source code is released under the GPL-3.0 license for acadmic purposes. The only requirement for using the code in your research is to cite our paper:

@article{Kang:2019:JOINT,
author = {Kang, Kaizhang and Xie, Cihui and He, Chengan and Yi, Mingqi and Gu, Minyi and Chen, Zimin and Zhou, Kun and Wu, Hongzhi},
title = {Learning Efficient Illumination Multiplexing for Joint Capture of Reflectance and Shape},
journal = {ACM Trans. Graph.},
issue_date = {November 2019},
volume = {38},
number = {6},
month = nov,
year = {2019},
issn = {0730-0301},
pages = {165:1--165:12},
articleno = {165},
numpages = {12},
url = {http://doi.acm.org/10.1145/3355089.3356492},
doi = {10.1145/3355089.3356492},
acmid = {3356492},
publisher = {ACM},
address = {New York, NY, USA},
keywords = {SVBRDF, lighting patterns, multi-view stereo, optimal sampling},
} 

For commercial licensing options, please email hwu at acm.org.
See COPYING for the open source license.

You can’t perform that action at this time.