A Multi-Head Approach with Shuffled Segments for Weakly-Supervised Video Anomaly Detection

Weakly-supervised video anomaly detection (WS-VAD) is a challenging task because coarse video-level annotations are insufficient to train fine-grained (segment or frame-level) detection algorithms. Multiple instance learning (MIL) powered by a ranking loss between the highest scoring segments of normal and anomaly videos has become the de-facto standard for WS-VAD. However, ranking loss is not robust to noisy segment-level labels (induced from the video-level labels), which is inherently the case in WS settings. In this work, we propose a new variant of the MIL method that utilizes a margin loss to achieve WS-VAD. The margin loss enables effective training of an anomaly scoring head based on noisy segment-level labels with high data imbalance (large number of normal segments and very few anomalous segments). We also introduce a self-supervised learning paradigm via stochastic shuffling of segments from multiple videos to mimic event changes during training. This forces the model to learn the boundaries between different virtual events (through a boundary localization head) and localizing the center of virtual events (through a center localization head). The efficacy of the proposed multi-head approach in successfully localizing anomalies is demonstrated through experiments on two large-scale VAD datasets (UCF-Crime and XD-Violence).

• Publication year

  2024

• Venue

  2024 IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW)

• Publication date

  2024-01-01

• Fields of study

  Computer Science

• Identifiers
  DOI 10.1109/WACVW60836.2024.00022
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Audiovisual Dependency Attention for Violence Detection in Videos

  2023cited by this paper
• PseudoBound: Limiting the anomaly reconstruction capability of one-class classifiers using pseudo anomalies

  2023cited by this paper
• Consistency-based Self-supervised Learning for Temporal Anomaly Localization

  2022cited by this paper
• A multi-stream deep neural network with late fuzzy fusion for real-world anomaly detection

  2022cited by this paper
• Modality-aware Contrastive Instance Learning with Self-Distillation for Weakly-Supervised Audio-Visual Violence Detection

  2022cited by this paper
• Self-supervised Sparse Representation for Video Anomaly Detection

  2022cited by this paper
• MGFN: Magnitude-Contrastive Glance-and-Focus Network for Weakly-Supervised Video Anomaly Detection

  2022influential reference
• Clustering Aided Weakly Supervised Training to Detect Anomalous Events in Surveillance Videos

  2022cited by this paper
• Generative Cooperative Learning for Unsupervised Video Anomaly Detection

  2022influential reference
• Real-World Video Anomaly Detection by Extracting Salient Features in Videos

  2022influential reference
• Look, Listen and Pay More Attention: Fusing Multi-Modal Information for Video Violence Detection

  2022cited by this paper
• Weakly-supervised Video Anomaly Detection with Robust Temporal Feature Magnitude Learning

  2021influential reference
• Violence Detection in Videos Based on Fusing Visual and Audio Information

  2021cited by this paper
• Anomaly Detection in Video Sequences: A Benchmark and Computational Model

  2021cited by this paper
• Anomalous Event Recognition in Videos Based on Joint Learning of Motion and Appearance with Multiple Ranking Measures

  2021cited by this paper
• Cleaning Label Noise with Clusters for Minimally Supervised Anomaly Detection

  2021cited by this paper
• MIST: Multiple Instance Self-Training Framework for Video Anomaly Detection

  2021influential reference
• Learning Causal Temporal Relation and Feature Discrimination for Anomaly Detection

  2021cited by this paper
• Multiple Instance-Based Video Anomaly Detection using Deep Temporal Encoding-Decoding

  2020cited by this paper
• Localizing Anomalies From Weakly-Labeled Videos

  2020cited by this paper
• Large-scale Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classification

  2020influential reference
• Not only Look, but also Listen: Learning Multimodal Violence Detection under Weak Supervision

  2020influential reference
• Video Anomaly Detection and Localization Based on an Adaptive Intra-Frame Classification Network

  2020cited by this paper
• DSNet: A Flexible Detect-to-Summarize Network for Video Summarization

  2020cited by this paper
• CLAWS: Clustering Assisted Weakly Supervised Learning with Normalcy Suppression for Anomalous Event Detection

  2020influential reference
• Anomaly Detection in Video via Self-Supervised and Multi-Task Learning

  2020cited by this paper
• Exploring Background-bias for Anomaly Detection in Surveillance Videos

  2019cited by this paper
• Generalization of feature embeddings transferred from different video anomaly detection domains

  2019cited by this paper
• Anomaly Locality in Video Surveillance

  2019cited by this paper
• Graph Convolutional Label Noise Cleaner: Train a Plug-And-Play Action Classifier for Anomaly Detection

  2019cited by this paper
• AnomalyNet: An Anomaly Detection Network for Video Surveillance

  2019cited by this paper
• Memorizing Normality to Detect Anomaly: Memory-Augmented Deep Autoencoder for Unsupervised Anomaly Detection

  2019cited by this paper
• Motion-Aware Feature for Improved Video Anomaly Detection

  2019cited by this paper
• Deep Self-Learning From Noisy Labels

  2019cited by this paper
• Deep and Sparse features For Anomaly Detection and Localization in video

  2019cited by this paper
• Temporal Convolutional Network with Complementary Inner Bag Loss for Weakly Supervised Anomaly Detection

  2019cited by this paper
• Social MIL: Interaction-Aware for Crowd Anomaly Detection

  2019cited by this paper
• 3D ResNet with Ranking Loss Function for Abnormal Activity Detection in Videos

  2019cited by this paper
• Dynamic video anomaly detection and localization using sparse denoising autoencoders

  2018cited by this paper
• Real-World Anomaly Detection in Surveillance Videos

  2018influential reference
• Object-Centric Auto-Encoders and Dummy Anomalies for Abnormal Event Detection in Video

  2018cited by this paper
• Video Jigsaw: Unsupervised Learning of Spatiotemporal Context for Video Action Recognition

  2018cited by this paper
• A study of deep convolutional auto-encoders for anomaly detection in videos

  2018cited by this paper
• Multiple-Instance Learning for Medical Image and Video Analysis

  2017cited by this paper
• Deep-Cascade: Cascading 3D Deep Neural Networks for Fast Anomaly Detection and Localization in Crowded Scenes

  2017cited by this paper
• Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset

  2017cited by this paper
• Learning from Noisy Labels with Distillation

  2017cited by this paper
• Spatio-Temporal AutoEncoder for Video Anomaly Detection

  2017cited by this paper
• Fully Convolutional Neural Network for Fast Anomaly Detection in Crowded Scenes

  2016cited by this paper
• Noise Versus Outliers

  2016cited by this paper
• Video anomaly detection and localisation based on the sparsity and reconstruction error of auto-encoder

  2016cited by this paper
• Making Deep Neural Networks Robust to Label Noise: A Loss Correction Approach

  2016cited by this paper
• Video anomaly detection using deep incremental slow feature analysis network

  2016cited by this paper
• Anomaly Detection in Video Using Predictive Convolutional Long Short-Term Memory Networks

  2016cited by this paper
• Learning Temporal Regularity in Video Sequences

  2016cited by this paper
• Practical Machine Learning: A New Look at Anomaly Detection

  2014cited by this paper
• Action Recognition

  2014cited by this paper
• Abnormal Event Detection at 150 FPS in MATLAB

  2013cited by this paper
• Multiple instance learning with bag dissimilarities

  2013cited by this paper
• Learning with Noisy Labels

  2013cited by this paper
• Mining top-n local outliers in large databases

  2001cited by this paper
• A Framework for Multiple-Instance Learning

  1997cited by this paper

• Weakly Supervised Video Anomaly Detection Based on Spatiotemporal Dual-Memory Enhancement and Contrastive Magnitude

  2025influential citation
• RelVid: Relational Learning with Vision-Language Models for Weakly Video Anomaly Detection

  2025cites this paper
• Multilingual-Prompt-Guided Directional Feature Learning for Weakly Supervised Video Anomaly Detection

  2025cites this paper
• The Evolution of Video Anomaly Detection: A Unified Framework from DNN to MLLM

  2025cites this paper
• StreamVAD: A streaming framework with progressive context integration for multi-temporal scale video anomaly detection

  2025cites this paper
• Deep Learning for Video Anomaly Detection: A Review

  2024cites this paper
• Cross-Modal Fusion and Attention Mechanism for Weakly Supervised Video Anomaly Detection

  2024influential citation
• A Hybrid U-Net and Vision Transformer approach for Video Anomaly detection

  2024cites this paper
• Learning Prompt-Enhanced Context Features for Weakly-Supervised Video Anomaly Detection

  2023cites this paper