Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Motion Semantics Guided Normalizing Flow for Privacy-Preserving Video Anomaly Detection

Yang Liu, Boan Chen, Yuanyuan Meng, Jing Liu, Zhengliang Guo, Wei Zhou, Peng Sun, Hong Chen

Abstract

As embodied perception systems increasingly bridge digital and physical realms in interactive multimedia applications, the need for privacy-preserving approaches to understand human activities in physical environments has become paramount. Video anomaly detection is a critical task in such embodied multimedia systems for intelligent surveillance and forensic analysis. Skeleton-based approaches have emerged as a privacy-preserving alternative that processes physical world information through abstract human pose representations while discarding sensitive visual attributes such as identity and facial features. However, existing skeleton-based methods predominantly model continuous motion trajectories in a monolithic manner, failing to capture the hierarchical nature of human activities composed of discrete semantic primitives and fine-grained kinematic details, which leads to reduced discriminability when anomalies manifest at different abstraction levels. In this regard, we propose Motion Semantics Guided Normalizing Flow (MSG-Flow) that decomposes skeleton-based VAD into hierarchical motion semantics modeling. It employs vector quantized variational auto-encoder to discretize continuous motion into interpretable primitives, an autoregressive Transformer to model semantic-level temporal dependencies, and a conditional normalizing flow to capture detail-level pose variations. Extensive experiments on benchmarks (HR-ShanghaiTech & HR-UBnormal) demonstrate that MSG-Flow achieves state-of-the-art performance with 88.1% and 75.8% AUC respectively.

Motion Semantics Guided Normalizing Flow for Privacy-Preserving Video Anomaly Detection

Abstract

As embodied perception systems increasingly bridge digital and physical realms in interactive multimedia applications, the need for privacy-preserving approaches to understand human activities in physical environments has become paramount. Video anomaly detection is a critical task in such embodied multimedia systems for intelligent surveillance and forensic analysis. Skeleton-based approaches have emerged as a privacy-preserving alternative that processes physical world information through abstract human pose representations while discarding sensitive visual attributes such as identity and facial features. However, existing skeleton-based methods predominantly model continuous motion trajectories in a monolithic manner, failing to capture the hierarchical nature of human activities composed of discrete semantic primitives and fine-grained kinematic details, which leads to reduced discriminability when anomalies manifest at different abstraction levels. In this regard, we propose Motion Semantics Guided Normalizing Flow (MSG-Flow) that decomposes skeleton-based VAD into hierarchical motion semantics modeling. It employs vector quantized variational auto-encoder to discretize continuous motion into interpretable primitives, an autoregressive Transformer to model semantic-level temporal dependencies, and a conditional normalizing flow to capture detail-level pose variations. Extensive experiments on benchmarks (HR-ShanghaiTech & HR-UBnormal) demonstrate that MSG-Flow achieves state-of-the-art performance with 88.1% and 75.8% AUC respectively.

Paper Structure

This paper contains 19 sections, 12 equations, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overview
  4. Motion Primitive Learning via Vector Quantization
  5. Motion Semantics Modeling with Primitive Flow
  6. Fine-Grained Variations Modeling with Detail Flow
  7. Training Objectives and Strategy
  8. Anomaly Detection
  9. Experiments
  10. Experimental Setup
  11. Datasets
  12. Evaluation Metrics
  13. Implementation Details
  14. Comparison with State-of-the-Art Methods
  15. Ablation Studies
  16. ...and 4 more sections

Figures (2)

  • Figure 1: Overview of MSG-Flow framework. The input skeleton sequence is encoded into discrete motion primitives through VQ-VAE. The primitive sequence is modeled by an autoregressive Transformer (Primitive Flow), while residual details are modeled by a conditional normalizing flow (Detail Flow).
  • Figure 2: Sensitivity analysis on (a) Codebook size $K$, (b) Window length $T$, and (c) Flow steps $K_{\text{flow}}$. Solid/dashed lines: HR-STC/HR-UBnormal.