Wen Liu*, Weixin Luo*, Zhengxin Li, Peilin Zhao, Shenghua Gao.
pip install -r requirements.txtPlease manually download all datasets from avenue.tar.gz and shanghaitech.tar.gz and tar each tar.gz file, and move them in to data folder.
You can also download data from BaiduYun(https://pan.baidu.com/s/1j0TEt-2Dw3kcfdX-LCF0YQ) i9b3
Download the pre-trained models firstly, pretrains folder
and then, move the pretrains folder into data,mv pretrains data.
python inference.py --dataset avenue \
--prednet cyclegan_convlstm \
--num_his 4 \
--label_level normal \
--gpu 0 \
--interpolation --snapshot_dir ./data/pretrains/avenue/normal/checkpoints/model.ckpt-74000
python inference.py --dataset avenue \
--prednet cyclegan_convlstm \
--num_his 4 \
--label_level tune_video \
--gpu 0 \
--interpolation --snapshot_dir ./data/pretrains/avenue/tune_video/prednet_cyclegan_convlstm_folds_10_kth_1_/MARGIN_1.0_LAMBDA_1.0/model.ckpt-76000
python inference.py --dataset avenue \
--prednet cyclegan_convlstm \
--num_his 4 \
--label_level normal \
--gpu 0 \
--interpolation --snapshot_dir ./data/pretrains/avenue/temporal/prednet_cyclegan_convlstm_folds_10_kth_1_/MARGIN_1.0_LAMBDA_1.0/model.ckpt-77000
See more details in
4.1 only_normal_data;
4.2 video_annotation;
4.3 temporal_annotation.
@inproceedings{melp_2019,
author = {Wen Liu and
Weixin Luo and
Zhengxin Li and
Peilin Zhao and
Shenghua Gao},
title = {Margin Learning Embedded Prediction for Video Anomaly Detection with
{A} Few Anomalies},
booktitle = {Proceedings of the Twenty-Eighth International Joint Conference on
Artificial Intelligence, {IJCAI} 2019, Macao, China, August 10-16,
2019},
pages = {3023--3030},
publisher = {ijcai.org},
year = {2019}
}