3D-Convolutional Neural Network with Generative Adversarial Network and Autoencoder for Robust Anomaly Detection in Video Surveillance

Int J Neural Syst. 2020 Jun;30(6):2050034. doi: 10.1142/S0129065720500343. Epub 2020 May 28.


As the surveillance devices proliferate, various machine learning approaches for video anomaly detection have been attempted. We propose a hybrid deep learning model composed of a video feature extractor trained by generative adversarial network with deficient anomaly data and an anomaly detector boosted by transferring the extractor. Experiments with UCSD pedestrian dataset show that it achieves 94.4% recall and 86.4% precision, which is the competitive performance in video anomaly detection.

Keywords: 3D CNN; Video anomaly detection; generative adversarial network; machine learning; transfer learning.

MeSH terms

  • Adult
  • Deep Learning* / standards
  • Humans
  • Pattern Recognition, Automated* / standards
  • Sensitivity and Specificity
  • Signal Detection, Psychological
  • Transfer, Psychology*
  • Video Recording