Home

Awesome

FedVAD: Enhancing Federated Video Anomaly Detection with GPT-Driven Semantic Distillation

Introduction

This repo is the official implementation of "FedVAD: Enhancing Federated Video Anomaly Detection with GPT-Driven Semantic Distillation". image

Requirements

The code requires python>=3.8 and the following packages:

torch==1.8.0
torchvision==0.9.0
numpy==1.21.2
scikit-learn==1.0.1
scipy==1.7.2
pandas==1.3.4
tqdm==4.63.0
xlwt==2.5

The environment with required packages can be created directly by running the following command:

conda env create -f environment.yml

Datasets

For the UCF-Crime and XD-Violence datasets, we use off-the-shelf features extracted by Wu et al. For the ShanghaiTech dataset, we used this repo to extract I3D features (highly recommended:+1:).

DatasetOrigin VideoI3D Features
  UCF-Crime  homepagedownload link
 XD-Violence  homepagedownload link
ShanghaiTech  homepagedownload link
   UBnormal  homepagedownload link

Acknowledgement

Our codebase mainly refers to STG-NF and PEL4VAD. We greatly appreciate their excellent contribution with nicely organized code!