/AutoregressModel-AE_VAD_CVPR2019

AutoregressModel-AE_VAD_CVPR2019 (code reimplemetation)

Primary LanguagePythonMIT LicenseMIT

AutoregressModel-AE_VAD_CVPR2019

This repo is the pytorch re-implemention of Latent Space Autoregression for Novelty Detection , CVPR 2019 by Davide Abati and Rita Cucchiara. The official imcompleted_pytorch code is here: official imcompleted_pytorch implementation

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

2. Download datasets

3. Testing on saved models

4. Training from scratch (here we use ped2 and avenue datasets for examples)

Notes

Citation