Detect anomaly on test video

vt-le / astnet

This is an official implementation for "Attention-based Residual Autoencoder for Video Anomaly Detection".

https://vt-le.github.io/astnet/

MIT License

102 stars 15 forks source link

Detect anomaly on test video #4

Closed Forchapeatl closed 1 year ago

Forchapeatl commented 2 years ago

Thanks for the code , how can one detect anomaly on test video

vt-le commented 2 years ago

Hi, thank you for your interest! You can follow the guide (here). There are several steps:

Clone the project
Download an anomaly dataset and pretrained model
Prepare your dataset

Forchapeatl commented 2 years ago

@vt-le , thank you. Please how my i visuaize the results on sample video

I have gone through the config file. Please how may i pass file "video.mp4" as parameter ?

vt-le commented 2 years ago

@vt-le , thank you. Please how my i visuaize the results on sample video

I have gone through the config file. Please how may i pass file "video.mp4" as parameter ?

Hi,

In this repository, we don't provide the code for visualization.
You cannot pass a name a video as the input. Only a name of a anomaly detection dataset [Ped2 / Avenue / ShanghaiTech] is accepted in config file.

Forchapeatl commented 2 years ago

Thank you

hhhh1230kvj commented 1 year ago

Thank you for your work, I just started this field, could you please send me your visualization code, thank you very much. 1418319084@qq.com.

lq-zhu commented 1 year ago

@vt-le Thanks for the code. I have finished all the preparations, but it's failed when test the dataset by running the following command. The "file" in "--cfg /path/to/config/file" confused me. If I want to test the avenue dataset, can you provide me with the correct run command?

Hi, thank you for your interest! You can follow the guide (here). There are several steps:

Clone the project

Download an anomaly dataset and pretrained model

Prepare your dataset

vt-le commented 1 year ago

@vt-le Thanks for the code. I have finished all the preparations, but it's failed when test the dataset by running the following command. The "file" in "--cfg /path/to/config/file" confused me. If I want to test the avenue dataset, can you provide me with the correct run command?

Hi, thank you for your interest! You can follow the guide (here). There are several steps:

Clone the project

Download an anomaly dataset and pretrained model

Prepare your dataset

Hi @lq-zhu , You can download the pre-trained model and save to pre_trained folder, then you can run like so: python astnet.py --cfg experiments/ave/ave_wresnet.yaml --model-file pre_trained/avenue.pth GPUS [1]