# Unsupervised Video Anomaly Detection via Normalizing Flows with Implicit Latent Features

arXiv: https://arxiv.org/abs/2010.07524 · HTML: https://mlsanity.com/paper/2010.07524/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 15 Oct 2020.

## Built on by (most cited first, 23 of 23)

- [Generative Cooperative Learning for Unsupervised Video Anomaly Detection](https://mlsanity.com/paper/2203.03962/) (2022)
- [Normalizing Flows for Human Pose Anomaly Detection](https://mlsanity.com/paper/2211.10946/) (2022)
- [Towards Open Set Video Anomaly Detection](https://mlsanity.com/paper/2208.11113/) (2022)
- [Cross-Domain Video Anomaly Detection without Target Domain Adaptation](https://mlsanity.com/paper/2212.07010/) (2022)
- [Deep Learning Advancements in Anomaly Detection: A Comprehensive Survey](https://mlsanity.com/paper/2503.13195/) (2025)
- [FastAno: Fast Anomaly Detection via Spatio-temporal Patch Transformation](https://mlsanity.com/paper/2106.08613/) (2021)
- [ProtoFL: Unsupervised Federated Learning via Prototypical Distillation](https://mlsanity.com/paper/2307.12450/) (2023)
- [Context-aware Video Anomaly Detection in Long-Term Datasets](https://mlsanity.com/paper/2404.07887/) (2024)
- [Event-Free Moving Object Segmentation from Moving Ego Vehicle](https://mlsanity.com/paper/2305.00126/) (2023)
- [Unsupervised anomaly detection for a Smart Autonomous Robotic Assistant Surgeon (SARAS)using a deep residual autoencoder](https://mlsanity.com/paper/2104.11008/) (2021)
- [Vision-Language Models Assisted Unsupervised Video Anomaly Detection](https://mlsanity.com/paper/2409.14109/) (2024)
- [Feature Density Estimation for Out-of-Distribution Detection via Normalizing Flows](https://mlsanity.com/paper/2402.06537/) (2024)
- [Detecting Contextual Anomalies by Discovering Consistent Spatial Regions](https://mlsanity.com/paper/2501.08470/) (2025)
- [RandomSEMO: Normality Learning Of Moving Objects For Video Anomaly Detection](https://mlsanity.com/paper/2202.06256/) (2022)
- [Clear Memory-Augmented Auto-Encoder for Surface Defect Detection](https://mlsanity.com/paper/2208.03879/) (2022)
- [Divide and Conquer in Video Anomaly Detection: A Comprehensive Review and New Approach](https://mlsanity.com/paper/2309.14622/) (2023)
- [Human Kinematics-inspired Skeleton-based Video Anomaly Detection](https://mlsanity.com/paper/2309.15662/) (2023)
- [EndoBoost: a plug-and-play module for false positive suppression during computer-aided polyp detection in real-world colonoscopy (with dataset)](https://mlsanity.com/paper/2212.12204/) (2022)
- [An optimization method for out-of-distribution anomaly detection models](https://mlsanity.com/paper/2302.00939/) (2023)
- [Out-of-distribution detection using normalizing flows on the data manifold](https://mlsanity.com/paper/2308.13792/) (2023)
- [Recognition of Abnormal Events in Surveillance Videos using Weakly Supervised Dual-Encoder Models](https://mlsanity.com/paper/2511.13276/) (2025)
- [Geometry-Aware Semantic Reasoning for Training Free Video Anomaly Detection](https://mlsanity.com/paper/2603.13374/) (2026)
- [Look Inside Each Video: Rethinking How Video Anomaly Detection Is Evaluated](https://mlsanity.com/paper/2608.21854/) (2026)

## Abstract (arXiv)

In contemporary society, surveillance anomaly detection, i.e., spotting anomalous events such as crimes or accidents in surveillance videos, is a critical task. As anomalies occur rarely, most training data consists of unlabeled videos without anomalous events, which makes the task challenging. Most existing methods use an autoencoder (AE) to learn to reconstruct normal videos; they then detect anomalies based on their failure to reconstruct the appearance of abnormal scenes. However, because anomalies are distinguished by appearance as well as motion, many previous approaches have explicitly separated appearance and motion information-for example, using a pre-trained optical flow model. This explicit separation restricts reciprocal representation capabilities between two types of information. In contrast, we propose an implicit two-path AE (ITAE), a structure in which two encoders implicitly model appearance and motion features, along with a single decoder that combines them to learn normal video patterns. For the complex distribution of normal scenes, we suggest normal density estimation of ITAE features through normalizing flow (NF)-based generative models to learn the tractable likelihoods and identify anomalies using out of distribution detection. NF models intensify ITAE performance by learning normality through implicitly learned features. Finally, we demonstrate the effectiveness of ITAE and its feature distribution modeling on six benchmarks, including databases that contain various anomalies in real-world scenarios.