NOLA

Repository for the NOLA dataset proposed in our paper Rethinking Video Anomaly Detection-A Continual Learning Approach.

It is implemented in tensorflow. Please follow the instructions to run the code.

1. Installation (Anaconda with python3.6 installation is recommended)

• Install 3rd-package dependencies of python (listed in requirements.txt)

numpy==1.14.1
scipy==1.0.0
matplotlib==2.1.2
tensorflow-gpu==1.4.1
tensorflow==1.4.1
Pillow==5.0.0
pypng==0.0.18
scikit_learn==0.19.1
opencv-python==3.2.0.6

pip install -r requirements.txt

pip install tensorflow-gpu==1.4.1

• Other libraries

CUDA 8.0
Cudnn 6.0
Ubuntu 14.04 or 16.04, Centos 7 and other distributions.

2. Download datasets

Please manually download the dataset from https://noladataset.s3.us-west-2.amazonaws.com/NOLA.zip and extract it, and move them in to Data folder.

3. Proposed Average Precision Delay Metric

4. Run the approach for a certain stage/split

python main.py $split
python main.py 1

Please note that you might need to adjust the knn threshold and number of epochs as you progress through the CL-stages since the available data size keeps on increasing. The parameters are already optimized for stage 1.