File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: A Causal Inference Look At Unsupervised Video Anomaly Detection
Title | A Causal Inference Look At Unsupervised Video Anomaly Detection |
---|---|
Authors | |
Issue Date | 2022 |
Publisher | AAAI Press. The Journal's web site is located at https://aaai.org/Library/AAAI/aaai-library.php |
Citation | 36th AAAI Conference on Artificial Intelligence (Virtual), February 22-March 1, 2022. In Proceedings of the AAAI Conference on Artificial Intelligence, v. 36 n. 2, p. 1620-1629 How to Cite? |
Abstract | Unsupervised video anomaly detection, a task that requires no labeled normal/abnormal training data in any form, is challenging yet of great importance to both industrial applications and academic research. Existing methods typically follow an iterative pseudo label generation process. However, they lack a principled analysis of the impact of such pseudo label generation on training. Furthermore, the long-range temporal dependencies also has been overlooked, which is unreasonable since the definition of an abnormal event depends on the long-range temporal context. To this end, first, we propose a causal graph to analyze the confounding effect of the pseudo label generation process. Then, we introduce a simple yet effective causal inference based framework to disentangle the noisy pseudo label's impact. Finally, we perform counterfactual based model ensemble that blends long-range temporal context with local image context in inference to make final anomaly detection. Extensive experiments on six standard benchmark datasets show that our proposed method significantly outperforms previous state-of-the-art methods, demonstrating our framework's effectiveness. |
Description | The conference program cochairs were Vasant Honavar (Pennsylvania State University, USA) and Matthijs Spaan (Delft University of Technology, Netherlands) AAAI Technical Track on Computer Vision II |
Persistent Identifier | http://hdl.handle.net/10722/316283 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Lin, X | - |
dc.contributor.author | Chen, Y | - |
dc.contributor.author | Li, G | - |
dc.contributor.author | Yu, Y | - |
dc.date.accessioned | 2022-09-02T06:08:45Z | - |
dc.date.available | 2022-09-02T06:08:45Z | - |
dc.date.issued | 2022 | - |
dc.identifier.citation | 36th AAAI Conference on Artificial Intelligence (Virtual), February 22-March 1, 2022. In Proceedings of the AAAI Conference on Artificial Intelligence, v. 36 n. 2, p. 1620-1629 | - |
dc.identifier.uri | http://hdl.handle.net/10722/316283 | - |
dc.description | The conference program cochairs were Vasant Honavar (Pennsylvania State University, USA) and Matthijs Spaan (Delft University of Technology, Netherlands) | - |
dc.description | AAAI Technical Track on Computer Vision II | - |
dc.description.abstract | Unsupervised video anomaly detection, a task that requires no labeled normal/abnormal training data in any form, is challenging yet of great importance to both industrial applications and academic research. Existing methods typically follow an iterative pseudo label generation process. However, they lack a principled analysis of the impact of such pseudo label generation on training. Furthermore, the long-range temporal dependencies also has been overlooked, which is unreasonable since the definition of an abnormal event depends on the long-range temporal context. To this end, first, we propose a causal graph to analyze the confounding effect of the pseudo label generation process. Then, we introduce a simple yet effective causal inference based framework to disentangle the noisy pseudo label's impact. Finally, we perform counterfactual based model ensemble that blends long-range temporal context with local image context in inference to make final anomaly detection. Extensive experiments on six standard benchmark datasets show that our proposed method significantly outperforms previous state-of-the-art methods, demonstrating our framework's effectiveness. | - |
dc.language | eng | - |
dc.publisher | AAAI Press. The Journal's web site is located at https://aaai.org/Library/AAAI/aaai-library.php | - |
dc.relation.ispartof | Proceedings of the AAAI Conference on Artificial Intelligence | - |
dc.title | A Causal Inference Look At Unsupervised Video Anomaly Detection | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Yu, Y: yzyu@cs.hku.hk | - |
dc.identifier.authority | Yu, Y=rp01415 | - |
dc.identifier.doi | 10.1609/aaai.v36i2.20053 | - |
dc.identifier.hkuros | 336339 | - |
dc.identifier.volume | 36 | - |
dc.identifier.issue | 2 | - |
dc.identifier.spage | 1620 | - |
dc.identifier.epage | 1629 | - |
dc.publisher.place | United States | - |