Video Anomaly Detection by Estimating Likelihood of Representations

TL;DR

This paper introduces a deep probabilistic model for video anomaly detection that estimates the likelihood of latent representations, effectively capturing spatial connections and improving detection performance on benchmark datasets.

Contribution

It proposes a novel deep denoising autoencoder combined with expectation maximization to model latent feature density for anomaly detection, addressing spatial connection limitations.

Findings

Achieves outstanding performance on benchmark datasets.

Effectively models latent feature density for anomaly detection.

Outperforms traditional clustering-based methods.

Abstract

Video anomaly detection is a challenging task not only because it involves solving many sub-tasks such as motion representation, object localization and action recognition, but also because it is commonly considered as an unsupervised learning problem that involves detecting outliers. Traditionally, solutions to this task have focused on the mapping between video frames and their low-dimensional features, while ignoring the spatial connections of those features. Recent solutions focus on analyzing these spatial connections by using hard clustering techniques, such as K-Means, or applying neural networks to map latent features to a general understanding, such as action attributes. In order to solve video anomaly in the latent feature space, we propose a deep probabilistic model to transfer this task into a density estimation problem where latent manifolds are generated by a deep denoising autoencoder and clustered by expectation maximization. Evaluations on several benchmarks datasets show the strengths of our model, achieving outstanding performance on challenging datasets.