https://vt-le.github.io/astnet/
This is an official implementation for "Attention-based Residual Autoencoder for Video Anomaly Detection".
https://vt-le.github.io/astnet/
anomaly-detection anomalydetection astnet avenue ped2 shanghaitech video-anomaly-detection videoanomalydetection
Last synced: 3 months ago
JSON representation
This is an official implementation for "Attention-based Residual Autoencoder for Video Anomaly Detection".
- Host: GitHub
- URL: https://vt-le.github.io/astnet/
- Owner: vt-le
- License: mit
- Created: 2021-12-15T12:19:27.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2025-01-17T22:25:48.000Z (10 months ago)
- Last Synced: 2025-08-09T04:31:45.280Z (3 months ago)
- Topics: anomaly-detection, anomalydetection, astnet, avenue, ped2, shanghaitech, video-anomaly-detection, videoanomalydetection
- Language: Python
- Homepage: https://vt-le.github.io/astnet/
- Size: 44.1 MB
- Stars: 123
- Watchers: 2
- Forks: 17
- Open Issues: 23
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Video-Anomaly-Detection - ASTNet: Attention-based Residual Autoencoder for Video Anomaly Detection
README
# ASTNet: Attention-based Residual Autoencoder for Video Anomaly Detection
This is the official implementation of **[Attention-based Residual Autoencoder for Video Anomaly Detection](https://doi.org/10.1007/s10489-022-03613-1)** [](https://hits.seeyoufarm.com).
## Related works
> **HSTforU**: See [HSTforU: Anomaly Detection in Aerial and Ground-based Videos with Hierarchical Spatio-Temporal Transformer for U-net
](https://vt-le.github.io/HSTforU/).
> **CrossAnomaly**: See [CrossAnomaly: A Contextual Cross-Modality Framework for Video Anomaly Detection](https://github.com/vt-le/CrossAnomaly).
### [Project](https://vt-le.github.io/astnet/) | [Video](https://youtu.be/XOzXwKVKX-Y) | [Paper](http://dx.doi.org/10.1007/s10489-022-03613-1)
## Updates
* [6/01/2023] Training script of [ASTNet](https://vt-le.github.io/astnet/) is released.
* [5/25/2022] [ASTNet](https://vt-le.github.io/astnet/) is available online.
* [4/21/2022] Code of [ASTNet](https://vt-le.github.io/astnet/) is released!
## Star History
[](https://star-history.com/#vt-le/astnet&Date)
## Prerequisites
* Linux or macOS
* Python 3
* PyTorch 1.7.0
## Setup
The code can be run with Python 3.6 and above.
Install the required packages:
pip install -r requirements.txt
Clone this repo:
git clone https://github.com/vt-le/astnet.git
cd ASTNet/ASTNet
## Data preparation
We evaluate `ASTNet` on:
* UCSD Ped2
* CUHK Avenue
* ShanghaiTech Campus
A dataset is a directory with the following structure:
```bash
$ tree data
ped2/avenue
├── training
│ └── frames
│ ├── ${video_1}$
│ │ ├── 000.jpg
│ │ ├── 001.jpg
│ │ └── ...
│ ├── ${video_2}$
│ │ ├── 00.jpg
│ │ └── ...
│ └── ...
├── testing
│ └── frames
│ ├── ${video_1}$
│ │ ├── 000.jpg
│ │ ├── 001.jpg
│ │ └── ...
│ ├── ${video_2}$
│ │ ├── 000.jpg
│ │ └── ...
│ └── ...
└── ped2/avenue.mat
shanghaitech
├── training
│ └── frames
│ ├── ${video_1}$
│ │ ├── 000.jpg
│ │ ├── 001.jpg
│ │ └── ...
│ ├── ${video_2}$
│ │ ├── 00.jpg
│ │ └── ...
│ └── ...
├── testing
│ └── frames
│ ├── ${video_1}$
│ │ ├── 000.jpg
│ │ ├── 001.jpg
│ │ └── ...
│ ├── ${video_2}$
│ │ ├── 000.jpg
│ │ └── ...
│ └── ...
└── test_frame_mask
├── 01_0014.npy
├── 01_0015.npy
└── ...
```
## Evaluation
Please first download the pre-trained model
Dataset | Pretrained Model |
-----------------|:--------:
UCSD Ped2 | [github][1] / [drive][4]
CUHK Avenue | [github][2] / [drive][5]
ShanghaiTech | [github][3] / [drive][6]
[1]: https://github.com/vt-le/storage/raw/wresnet/ped2.pth
[2]: https://github.com/vt-le/storage/raw/wresnet/avenue.pth
[3]: https://github.com/vt-le/storage/raw/wresnet/shanghai.pth
[4]: https://drive.google.com/file/d/1dmDVyAkI0FiEex3aEHDQMdgDFeBtH1Fo/view?usp=sharing
[5]: https://drive.google.com/file/d/1EtHqDiq3tDikgpXCm6LltWzgSnSvyH1U/view?usp=sharing
[6]: https://drive.google.com/file/d/1SwUGiwyhEUPIk8CTSS_5ZDjg3Idi7CJS/view?usp=sharing
To evaluate a pretrained `ASTNet` on a dataset, run:
```bash
python test.py \
--cfg \
--model-file
```
For example, to evaluate `ASTNet` on Ped2:
```bash
python test.py \
--cfg config/ped2_wresnet.yaml \
--model-file pretrained.ped2.pth
```
## Training from scratch
To train `ASTNet` on a dataset, run:
```bash
python train.py \
--cfg
```
For example, to train `ASTNet` on Ped2:
```bash
python train.py \
--cfg config/ped2_wresnet.yaml
```
**Notes**:
- To change other options, see ``.
## Citing
If you find our work useful for your research, please consider citing:
```BibTeX
@article{le2023attention,
title={Attention-based Residual Autoencoder for Video Anomaly Detection},
author={Le, Viet-Tuan and Kim, Yong-Guk},
journal={Applied Intelligence},
volume={53},
number={3},
pages={3240--3254},
year={2023},
publisher={Springer}
}
```
## Contact
For any question, please file an [issue](https://github.com/vt-le/astnet/issues) or contact:
Viet-Tuan Le: vt-le@outlook.com