csjfwang / st-cgan

Dataset and Code for our CVPR'18 paper ST-CGAN: "Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal"

Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal

ST-CGAN: [Arxiv]

by Jifeng Wang, Xiang Li and Jian Yang
DeepInsight@PCALab, Nanjing University of Science and Technology.

---

Citation

@inproceedings{wang2018STCGAN,
　author = {Wang, Jifeng and Li, Xiang and Yang, Jian},
　title = {Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal},
　booktitle = {CVPR},
　year = {2018}
}

ISTD Dataset

ISTD dataset is available in Google Drive

Removal Results on ISTD

Shadow removal results on ISTD is available in Google Drive (including the results of all the method we compared in our paper).

License

Copyright (c) 2018, Jifeng Wang All rights reserved.
The ISTD Dataset can only be used for research and Non-Commercial purposes.

Code

Coming soon...

st-cgan's Issues

UCF dataset cannot be accessed.Would you share it?

lots of thanks.

您好，关于源码问题。

最近看了相关的论文，也自己做了相关的实验，但是怀疑我写的评价指标部分的代码，和文章中所用有较大区别，所以想请问一下是否方便尽快开源一下评价指标那部分代码，想进行一下对比，谢谢您！

Test for research

Hello,
I see that this solution fit my research needs.
Currently i am testing only pix2pix variations like CGAN, GANBegin and another solution Photographic Image Synthesis with Cascaded Refinement Networks solutions. The main purpose is to inspect the quality of shadow removal by passing images with and without shadow.
As cGAN gave me the best results, but even those where not perfect for my training dataset, i would like to ask you for the code ? Just for testing purposes ! Is it possible ?