Dual-Space Normalizing Flow for Unsupervised Video Anomaly Detection
- 1 October 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 34 (10577149) , 6432-6445
Abstract
Conventional reconstruction-based video anomaly detection (VAD) methods implicitly model normality in latent spaces, which is limited by the generalization ability of latent features. Normalizing Flow (NF)-based methods have been introduced to address this issue, as they explicitly model the distribution of input data and achieve significant performance in VAD. However, existing NF-based methods are confined to Euclidean space, limiting their ability to model action hierarchies. While effective at capturing local joint dynamics and short-term temporal variations, they fail to encode kinematic dependencies and long-term pose evolution, ultimately struggling to discern ambiguous anomalies that deviate minimally from normal motion. In contrast, hyperbolic representation learning, with its ability to model hierarchical and complex relationships among actions, offers a promising solution to enhance the discriminative power between similar skeletal actions. Motivated by this, we propose a novel Dual-Space Normalizing Flow (DSNF) method. Specifically, we design a Dual-Space Parallel Graph Convolutional Network (DSPGCN) that synergistically integrates the strengths of both Euclidean and hyperbolic geometries to simultaneously capture local detail features of poses and intrinsic hierarchical relationships of actions. To enhance the model’s focus on discriminative features, we design an Adaptive Weighted Approximation Mass (AWAM) loss that dynamically adjusts weights to impose stronger constraints on regions with low discriminability in the dual space, encouraging the model to focus more on key discriminative features in hyperbolic space that reflect complex relationships between actions. Extensive experiments on public datasets demonstrate the effectiveness and robustness of our method in various VAD scenarios.Keywords
Funding Information
- National Key Research and Development Program of China (2022YFA1004100)
- National Natural Science Foundation of China (62472060, U23A20318, 62221005)
- Natural Science Foundation of Chongqing Municipality (CSTB2024NSCQ-QCXMX0060, CSTB2023NSCQ-LZX0061, CSTB2022NSCQ-MSX0547)
- Science and Technology Research Program of Chongqing Municipal Education Commission (KJZD-K202300604)
- Chongqing Institute for Brain and Intelligence
- China Postdoctoral Science Foundation (2025MD774186)
- Chongqing Special Postdoctoral Research Funding (2024CQBSHTB2002)
This publication has 49 references indexed in Scilit:
- Contracting skeletal kinematics for human-related video anomaly detectionPattern Recognition, 2024
- A Spatio‐Temporal Enhanced Graph‐Transformer AutoEncoder embedded pose for anomaly detectionIET Computer Vision, 2023
- Video Anomaly Detection by Solving Decoupled Spatio-Temporal Jigsaw Puzzles2022
- Skin Lesion Recognition with Class-Hierarchy Regularized Hyperbolic Embeddings2022
- Video anomaly detection using CycleGan based on skeleton featuresJournal of Visual Communication and Image Representation, 2022
- Feature Extraction Using an RNN Autoencoder for Skeleton-Based Abnormal Gait RecognitionIEEE Access, 2020
- Object-Centric Auto-Encoders and Dummy Anomalies for Abnormal Event Detection in Video2019
- CrowdPose: Efficient Crowded Scenes Pose Estimation and a New Benchmark2019
- Future Frame Prediction for Anomaly Detection - A New Baseline2018
- Learning Temporal Regularity in Video Sequences2016