Code and dataset for the CVPR 2021 paper "Triple-cooperative Video Shadow Detection"[arXiv link] [official link].
by Zhihao Chen¹, Liang Wan¹, Lei Zhu², Jia Shen¹, Huazhu Fu³, Wennan Liu⁴, and Jing Qin^5
1College of Intelligence and Computing, Tianjin University
²Department of Applied Mathematics and Theoretical Physics, University of Cambridge
³Inception Institute of Artificial Intelligence, UAE
⁴Academy of Medical Engineering and Translational Medicine, Tianjin University
⁵The Hong Kong Polytechnic University

@inproceedings{chen21TVSD,
     author = {Chen, Zhihao and Wan, Liang and Zhu, Lei and Shen, Jia and Fu, Huazhu and Liu, Wennan and Qin, Jing},
     title = {Triple-cooperative Video Shadow Detection},
     booktitle = {CVPR},
     year = {2021}
}

Pytorch-lightning Version is available at https://github.com/lihaoliu-cambridge/video-shadow-detection implemented by Lihao Liu

ViSha dataset is available at ViSha Homepage

• Python 3.6
• PyTorch 1.3.1
• torchvision
• numpy
• tqdm
• PIL
• math
• time
• datatime
• argparse
• apex (alternative, fp16 for save memory and speedup)

1. Modify the data path on ./config.py
2. Modify the pretrained backbone path on ./networks/resnext_modify/config.py
3. Run by python train.py and model will be saved in ./models/TVSD

The pretrained ResNeXt model is ported from the official torch version, using the convertor provided by clcarwin. You can directly download the pretrained model ported by us.

1. Modify the data path on ./config.py
2. Make sure you have a snapshot in ./models/TVSD (Tips: You can download the trained model which is reported in our paper at BaiduNetdisk(pw: 8p5h) or Google Drive)
3. Run by python infer.py to generate predicted masks
4. Run by python evaluate.py to evaluate the generated results

As mentioned in our paper, since there is no CNN-based method for video shadow detection, we make comparison against 12 state-of-the-art methods for relevant tasks, including BDRAR[1], DSD[2], MTMT[3] (single-image shadow detection), FPN[4], PSPNet[5] (single-image semantic segmentation), DSS[6], R^3 Net[7] (single-image saliency detection), PDBM[8], MAG[9] (video saliency detection), COSNet[10], FEELVOS[11], STM[12] (object object segmentation)
[1]L. Zhu, Z. Deng, X. Hu, C.-W. Fu, X. Xu, J. Qin, and P.-A. Heng. Bidirectional feature pyramid network with recurrent attention residual modules for shadow detection. In ECCV, pages 121–136, 2018.
[2]Q. Zheng, X. Qiao, Y. Cao, and R.W. Lau. Distraction-aware shadow detection. In CVPR, pages 5167–5176, 2019.
[3]Z. Chen, L. Zhu, L. Wan, S. Wang, W. Feng, and P.-A. Heng. A multi-task mean teacher for semi-supervised shadow detection. In CVPR, pages 5611–5620, 2020.
[4]T.-Y. Lin, P. Doll´ar, R. Girshick, K. He, B. Hariharan, and S.Belongie. Feature pyramid networks for object detection. In CVPR, pages 2117–2125, 2017.
[5]H. Zhao, J. Shi, X. Qi, X. Wang, and J. Jia. Pyramid scene parsing network. In CVPR, pages 2881–2890, 2017.
[6]Q. Hou, M. Cheng, X. Hu, A. Borji, Z. Tu, and P. Torr. Deeply supervised salient object detection with short connections. IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(4):815–828, 2019.
[7]Z. Deng, X. Hu, L. Zhu, X. Xu, J. Qin, G. Han, and P.-A. Heng. R3net: Recurrent residual refinement network for saliency detection. In IJCAI, pages 684–690. AAAI Press, 2018.
[8]H. Song, W. Wang, S. Zhao, J. Shen, and K.-M. Lam. Pyramid dilated deeper convlstm for video salient object detection. In ECCV, pages 715–731, 2018.
[9]H. Li, G. Chen, G. Li, and Y. Yu. Motion guided attention for video salient object detection. In ICCV, pages 7274–7283, 2019.
[10]X. Lu, W. Wang, C. Ma, J. Shen, L. Shao, and F. Porikli. See more, know more: Unsupervised video object segmentation with co-attention siamese networks. In CVPR, pages 3623–3632, 2019.
[11]P. Voigtlaender, Y. Chai, F. Schroff, H. Adam, B. Leibe, and L.-C. Chen. Feelvos: Fast end-to-end embedding learning for video object segmentation. In CVPR, June 2019.
[12]S.W. Oh, J.-Y. Lee, N. Xu, and S.J. Kim. Video object segmentation using space-time memory networks. In ICCV, pages 9226–9235, 2019.

We evaluate those methods and our TVSD in ViSha testing set and release all results in BaiduNetdisk(pw: ritw) or Google Drive