Lightning fast video anomaly detection via multi-scale adversarial distillationShow others and affiliations
2024 (English)In: Computer Vision and Image Understanding, ISSN 1077-3142, E-ISSN 1090-235X, Vol. 247, article id 104074Article in journal (Refereed) Published
Abstract [en]
We propose a very fast frame-level model for anomaly detection in video, which learns to detect anomalies by distilling knowledge from multiple highly accurate object-level teacher models. To improve the fidelity of our student, we distill the low-resolution anomaly maps of the teachers by jointly applying standard and adversarial distillation, introducing an adversarial discriminator for each teacher to distinguish between target and generated anomaly maps. We conduct experiments on three benchmarks (Avenue, ShanghaiTech, UCSD Ped2), showing that our method is over 7 times faster than the fastest competing method, and between 28 and 62 times faster than object-centric models, while obtaining comparable results to recent methods. Our evaluation also indicates that our model achieves the best trade-off between speed and accuracy, due to its previously unheard-of speed of 1480 FPS. In addition, we carry out a comprehensive ablation study to justify our architectural design choices. Our code is freely available at: https://github.com/ristea/fast-aed.
Place, publisher, year, edition, pages
ACADEMIC PRESS INC ELSEVIER SCIENCE , 2024. Vol. 247, article id 104074
Keywords [en]
Abnormal event detection; Anomaly detection; Knowledge distillation; Neural networks; Transformers
National Category
Computer Vision and Learning Systems
Identifiers
URN: urn:nbn:se:liu:diva-206265DOI: 10.1016/j.cviu.2024.104074ISI: 001276294600001Scopus ID: 2-s2.0-85199047468OAI: oai:DiVA.org:liu-206265DiVA, id: diva2:1888955
Note
Funding Agencies|Romanian Ministry of Education and Research, CNCS-UEFISCDI [PN-III-P2-2.1-PED-2021-0195]
2024-08-142024-08-142026-06-22