SSMTL++: Revisiting self-supervised multi-task learning for video anomaly detection

Antonio Barbalau, Radu Tudor Ionescu, Mariana-Iuliana Georgescu, Jacob V. Dueholm, Bharathkumar Ramachandra, Kamal Nasrollahi, Fahad Shahbaz Khan, Thomas B. Moeslund, Mubarak Shah

Abstract: Highlights•We introduce new detection methods into a self-supervised multi-task learning method.•We introduce convolutional vision transformer blocks into the backbone architecture.•We study various new proxy tasks for video anomaly detection.