Skip to content

Pytorch Re-implement of ano_pre_cvpr2018, flownet2 / lite-flownet used.

Notifications You must be signed in to change notification settings

fjchange/pytorch_ano_pre

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

67 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

pytorch_ano_pre

Pytorch Re-implemention of ano_pre_cvpr2018, replace flownet2 with lite-flownet

img

Future Frame Prediction for Anomaly Detection -- A New Baseline, CVPR 2018

tensorflow_offical_implement

** This repo modify the normalization of the Regular Score, And replace flownetSD with lite-flownet ** AUC 85.6%+-0.1% of Avenue dataset

You can use FlowNet2SD Now, modify the code in train.py as the comment said.

img

1. requirement

  • pytorch >=0.4.1
  • tensorboardX (if you want)

2. preparation

  1. Download Dataset CUHK Avenue download_link, unzip in the path you want, and replace the path in train.py

  2. Download Lite-Flownet model, and replace the path in train.py

wget --timestamping http://content.sniklaus.com/github/pytorch-liteflownet/network-sintel.pytorch

** The quality of optical flow matters, it would be better if you finetune the liteflownet with FlyingChairsSDHom dataset**

if you want to use FlowNet2SD, you should download model form Nvidia/flownet2-pytorch, and replace the path in train.py

Flownet2SD

  1. replace all the modle_output_path and log_output_path to where you want in train.py

3. training

cd ano_pre

python train.py

4. evalute

replace the model_path and evaluate_name as you want

cd ano_pre

python evaluate.py

img

5. reference

If you find this useful, please cite the work as follows:

[1]  @INPROCEEDINGS{liu2018ano_pred, 
        author={W. Liu and W. Luo, D. Lian and S. Gao}, 
        booktitle={2018 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, 
        title={Future Frame Prediction for Anomaly Detection -- A New Baseline}, 
        year={2018}   
     }   
[2]  misc{pytorch_ano_pred,
          author = {Jia-Chang Feng},
          title = { A Reimplementation of {Ano_pred} Using {Pytorch}},
          year = {2019},
          howpublished = {\url{https://github.com/fjchange/pytorch_ano_pre}}    
    }
[3]  @inproceedings{Hui_CVPR_2018,
         author = {Tak-Wai Hui and Xiaoou Tang and Chen Change Loy},
         title = {{LiteFlowNet}: A Lightweight Convolutional Neural Network for Optical Flow Estimation},
         booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
         year = {2018}  
     }
[4]  @misc{pytorch-liteflownet,
         author = {Simon Niklaus},
         title = {A Reimplementation of {LiteFlowNet} Using {PyTorch}},
         year = {2019},
         howpublished = {\url{https://github.com/sniklaus/pytorch-liteflownet}}      
    }

About

Pytorch Re-implement of ano_pre_cvpr2018, flownet2 / lite-flownet used.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published