Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
153 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

MULDE: Multiscale Log-Density Estimation via Denoising Score Matching for Video Anomaly Detection (2403.14497v1)

Published 21 Mar 2024 in cs.CV

Abstract: We propose a novel approach to video anomaly detection: we treat feature vectors extracted from videos as realizations of a random variable with a fixed distribution and model this distribution with a neural network. This lets us estimate the likelihood of test videos and detect video anomalies by thresholding the likelihood estimates. We train our video anomaly detector using a modification of denoising score matching, a method that injects training data with noise to facilitate modeling its distribution. To eliminate hyperparameter selection, we model the distribution of noisy video features across a range of noise levels and introduce a regularizer that tends to align the models for different levels of noise. At test time, we combine anomaly indications at multiple noise scales with a Gaussian mixture model. Running our video anomaly detector induces minimal delays as inference requires merely extracting the features and forward-propagating them through a shallow neural network and a Gaussian mixture model. Our experiments on five popular video anomaly detection benchmarks demonstrate state-of-the-art performance, both in the object-centric and in the frame-centric setup.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (48)
  1. Ubnormal: New benchmark for supervised open-set video anomaly detection. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
  2. Ssmtl++: Revisiting self-supervised multi-task learning for video anomaly detection. Computer Vision and Image Understanding, 229, 2023.
  3. Appearance-motion memory consistency network for video anomaly detection. In Proc. of the Conference on Artificial Intelligence (AAAI), 2021.
  4. Clustering driven deep autoencoder for video anomaly detection. In Proc. of the European Conference on Computer Vision (ECCV), 2020.
  5. Histograms of oriented optical flow and binet-cauchy kernels on nonlinear dynamical systems for the recognition of human actions. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2009.
  6. Histograms of optical flow orientation and magnitude and entropy to detect anomalous events in videos. Transactions on Circuits and Systems for Video Technology, 27(3):673–682, 2016.
  7. Rmpe: Regional multi-person pose estimation. In Proc. of the International Conference on Computer Vision (ICCV), 2017.
  8. Convolutional transformer based dual discriminator generative adversarial networks for video anomaly detection. In Proc. of the International Conference on Multimedia, 2021.
  9. Multimodal motion conditioned diffusion model for skeleton-based video anomaly detection. In Proc. of the International Conference on Computer Vision (ICCV), 2023.
  10. Anomaly detection in video via self-supervised and multi-task learning. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021a.
  11. A background-agnostic framework with adversarial training for abnormal event detection in video. Transactions on Pattern Analysis and Machine Intelligence (PAMI), 44(9), 2021b.
  12. Memorizing normality to detect anomaly: Memory-augmented deep autoencoder for unsupervised anomaly detection. In Proc. of the International Conference on Computer Vision (ICCV), 2019.
  13. Learning temporal regularity in video sequences. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
  14. Normalizing flows for human pose anomaly detection. In Proc. of the International Conference on Computer Vision (ICCV), 2023.
  15. Flownet 2.0: Evolution of optical flow estimation with deep networks. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
  16. Object-centric auto-encoders and dummy anomalies for abnormal event detection in video. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019.
  17. The kinetics human action video dataset. arXiv preprint arXiv:1705.06950, 2017.
  18. Adam: A method for stochastic optimization. In Proc. of the International Conference on Learning Representations (ICLR), 2015.
  19. Why normalizing flows fail to detect out-of-distribution data. In Proc. of the Conference on Neural Information Processing Systems (NeurIPS), 2020.
  20. A tutorial on energy-based learning. Predicting Structured Data, 1, 2006.
  21. Stan: Spatio- temporal adversarial networks for abnormal event detection. In Proc. of the International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2018.
  22. Bman: Bidirectional multi-scale aggregation networks for abnormal event detection. Transactions on Image Processing, 29, 2019.
  23. Future frame prediction for anomaly detection–a new baseline. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
  24. A hybrid video anomaly detection framework via memory-augmented flow reconstruction and flow-guided frame prediction. In Proc. of the International Conference on Computer Vision (ICCV), 2021.
  25. Abnormal event detection at 150 fps in matlab. In Proc. of the International Conference on Computer Vision (ICCV), 2013.
  26. A revisit of sparse coding based anomaly detection in stacked rnn framework. In Proc. of the International Conference on Computer Vision (ICCV), 2017.
  27. Learning Normal Dynamics in Videos with Meta Prototype Network. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021.
  28. Anomaly detection in crowded scenes. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2010.
  29. Multiscale score matching for out-of-distribution detection. Proc. of the International Conference on Learning Representations (ICLR), 2021.
  30. Anomaly detection in video sequence with appearance-motion correspondence. In Proc. of the International Conference on Computer Vision (ICCV), 2019.
  31. Learning memory-guided normality for anomaly detection. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
  32. Learning transferable visual models from natural language supervision. In Proc. of the International Conference on Machine Learning (ICML), 2021.
  33. Street scene: A new dataset and evaluation protocol for video anomaly detection. In Proc. of the Winter Conference on Applications of Computer Vision (WACV), 2020.
  34. Learning a distance function with a siamese network to localize anomalies in videos. In Proc. of the Winter Conference on Applications of Computer Vision (WACV), 2020.
  35. Attribute-based representations for accurate and interpretable video anomaly detection. arXiv preprint arXiv:2212.00789, 2022.
  36. Self-supervised predictive convolutional attentive block for anomaly detection. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
  37. Hiera: A hierarchical vision transformer without the bells-and-whistles. Proc. of the International Conference on Machine Learning (ICML), 2023.
  38. Generative modeling by estimating gradients of the data distribution. In Proc. of the Conference on Neural Information Processing Systems (NeurIPS), 2019.
  39. Real-world anomaly detection in surveillance videos. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
  40. Scene-aware context reasoning for unsupervised abnormal event detection in videos. In Proc. of the International Conference on Multimedia, 2020.
  41. Pascal Vincent. A connection between score matching and denoising autoencoders. Neural Computation, 23(7):1661–1674, 2011.
  42. Video anomaly detection by solving decoupled spatio-temporal jigsaw puzzles. In Proc. of the European Conference on Computer Vision (ECCV), 2022.
  43. Gods: Generalized one-class discriminative subspaces for anomaly detection. In Proc. of the International Conference on Computer Vision (ICCV), 2019.
  44. Cluster attention contrast for video anomaly detection. In Proc. of the International Conference on Multimedia, 2020.
  45. Feature prediction diffusion model for video anomaly detection. In Proc. of the International Conference on Computer Vision (ICCV), 2023.
  46. Dynamic local aggregation network with adaptive clusterer for anomaly detection. In Proc. of the European Conference on Computer Vision (ECCV), 2022.
  47. Cloze test helps: Effective video anomaly detection via learning to complete video events. In Proc. of the International Conference on Multimedia, 2020.
  48. Generative cooperative learning for unsupervised video anomaly detection. In Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now