• What we'll do
We will present two frameworks for abnormal event detection in video that require no training sequences. The first framework  detects changes on a sequence of data from the video to see which frames are distinguishable from all the previous frames. As the authors want to build an approach independent of temporal ordering, they create shuffles of the data by permuting the frames before running each instance of the change detection. The second framework  is based on unmasking, a technique previously used for authorship verification in text documents, which we adapted to abnormal event detection as follows: a binary classifier is iteratively trained to distinguish between two consecutive video sequences while removing at each step the most discriminant features. Higher training accuracy rates of the intermediately obtained classifiers represent abnormal events.
 ECCV 2016: https://www.ri.cmu.edu/pub_files/2016/10/anomalyframework.pdf
 ICCV 2017: http://openaccess.thecvf.com/content_ICCV_2017/papers/Ionescu_Unmasking_the_Abnormal_ICCV_2017_paper.pdf
Speaker: Radu Ionescu
Location: Stoilow hall, 1st floor, Faculty of Mathematics and Computer Science
Claim the event and start manage its content.I am the organizer