S2VD
所属分类:人工智能/神经网络/深度学习
开发工具:Python
文件大小:0KB
下载次数:0
上传日期:2021-05-29 08:01:12
上 传 者:
sh-1993
说明: 使用动态雨水发生器的半监督视频去训练(CVPR,2021,Pytorch),
(Semi-supervised Video Deraining with Dynamical Rain Generator (CVPR, 2021, Pytorch),)
文件列表:
LICENSE (1070, 2021-05-29)
S2VD-supp.pdf (1756841, 2021-05-29)
environment.yml (2294, 2021-05-29)
main_NTURain.py (579, 2021-05-29)
makedata/ (0, 2021-05-29)
makedata/preparedata_NTU.py (4098, 2021-05-29)
makedata/preparedata_NTU_semi.py (2719, 2021-05-29)
model_states/ (0, 2021-05-29)
model_states/derainer_rho05.pt (2149309, 2021-05-29)
networks/ (0, 2021-05-29)
networks/__init__.py (92, 2021-05-29)
networks/__pycache__/ (0, 2021-05-29)
networks/__pycache__/__init__.cpython-36.pyc (133, 2021-05-29)
networks/__pycache__/__init__.cpython-37.pyc (146, 2021-05-29)
networks/__pycache__/derain_net.cpython-36.pyc (2451, 2021-05-29)
networks/__pycache__/derain_net.cpython-37.pyc (2422, 2021-05-29)
networks/__pycache__/derain_net2.cpython-36.pyc (2369, 2021-05-29)
networks/__pycache__/derain_net3.cpython-36.pyc (2501, 2021-05-29)
networks/__pycache__/generators.cpython-36.pyc (3635, 2021-05-29)
networks/__pycache__/sub_blocks.cpython-36.pyc (3881, 2021-05-29)
networks/__pycache__/sub_blocks.cpython-37.pyc (3885, 2021-05-29)
networks/derain_net.py (3695, 2021-05-29)
networks/generators.py (5010, 2021-05-29)
networks/sub_blocks.py (3329, 2021-05-29)
options_derain.json (2118, 2021-05-29)
test_MSCSC_real.py (2088, 2021-05-29)
test_NTURain_real.py (2781, 2021-05-29)
test_NTURain_synthetic.py (3538, 2021-05-29)
testsets/ (0, 2021-05-29)
testsets/real_MSCSC/ (0, 2021-05-29)
testsets/real_MSCSC/Manhattan.mat (21581147, 2021-05-29)
testsets/real_MSCSC/Saigon.mat (38258422, 2021-05-29)
testsets/real_MSCSC/night.mat (16579472, 2021-05-29)
testsets/real_MSCSC/street.mat (29659954, 2021-05-29)
testsets/real_MSCSC/tokyo.mat (13118143, 2021-05-29)
testsets/real_NTURain/ (0, 2021-05-29)
testsets/real_NTURain/ra1_Rain/ (0, 2021-05-29)
testsets/real_NTURain/ra1_Rain/00001.jpg (57393, 2021-05-29)
testsets/real_NTURain/ra1_Rain/00002.jpg (56810, 2021-05-29)
... ...
# S2VD
# [Semi-supervised Video Deraining with Dynamical Rain Generator (CVPR, 2021)](https://arxiv.org/abs/2103.07939)
# Requirements and Dependencies
* Ubuntu 16.04, cuda 10.0
* Python 3.6.10, Pytorch 1.6.0
* More detail (See [environment.yml](environment.yml))
# Training pipelines
1. Download the NTURain dataset from [here](https://github.com/hotndy/SPAC-SupplementaryMaterials) or [Baidu Cloud](https://pan.baidu.com/s/1MrIU8RFedfw2ptuuVtHYVA)(Passwd:dtgv), and prepare the training data as follows:
- Labled synthetic data:
```python
python makedata/preparedata_NTU.py --ntu_path your_downloaded_synthetic_path --train_path your_saved_train_path
```
- Unlabled real data:
```python
python makedata/preparedata_NTU_semi.py --ntu_path_semi your_downloaded_real_path --train_path your_saved_train_path
```
Note that you should better put the synthetic and real training data sets into two different training folders.
2. Modify the configured file [options_derain.json](options_derain.json) according to your own training and testing path.
3. Begin training:
```
python main_NTURain.py
```
# Testing pipelines
You need firstly download the testing dataset of [NTURain](https://github.com/hotndy/SPAC-SupplementaryMaterials) and [MSCSC](MSCS://github.com/MinghanLi/MS-CSC-Rain-Streak-Removal) into the folder [testsets](testsets).
+ NTURain synthetic data set:
```
python test_NTURain_synthetic.py
```
This manuscript will re-produce the paper results in Table 1.
+ NTURain real data set:
```
python test_NTURain_real.py
```
+ MSCSC real data set:
```
python test_MSCSC_real.py
```
# Citation
```
@incollection{CVPR2021_2429,
title = {Semi-supervised video deraining with dynamical rain generator},
author = {Yue, Zongsheng and Xie, Jianwen and Zhao, Qian and Meng, Deyu},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2021}
}
```
近期下载者:
相关文件:
收藏者: