Towards Adaptive Human-centric Video Anomaly Detection: A Comprehensive Framework and A New Benchmark

2024-08-26

AI Summary

Human-centered video anomaly detection (VAD) faces challenges including high behavioral diversity, extreme scarcity of abnormal samples, and stringent privacy-ethics constraints—hindering scalable dataset construction and continual learning. To address these, we propose HuVAD, the first privacy-enhanced VAD benchmark, covering seven real-world scenarios and featuring a >5× increase in pose-annotated frames. We further introduce UCAL, an unsupervised continual anomaly learning framework integrating: (i) human-pose-driven feature modeling, (ii) differential-privacy-inspired de-identification for annotation, and (iii) a multi-granularity anomaly scoring mechanism enabling label-free incremental adaptation. Evaluated on both standard and continual-learning benchmarks, UCAL achieves state-of-the-art performance on 82.14% of metrics, demonstrating substantial improvements in long-tail anomaly recognition and cross-scenario generalization.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Machine Learning: Calibration & Uncertainty Quantification
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Human-centric Video Anomaly Detection (VAD) aims to identify human behaviors that deviate from normal. At its core, human-centric VAD faces substantial challenges, such as the complexity of diverse human behaviors, the rarity of anomalies, and ethical constraints. These challenges limit access to high-quality datasets and highlight the need for a dataset and framework supporting continual learning. Moving towards adaptive human-centric VAD, we introduce the HuVAD (Human-centric privacy-enhanced Video Anomaly Detection) dataset and a novel Unsupervised Continual Anomaly Learning (UCAL) framework. UCAL enables incremental learning, allowing models to adapt over time, bridging traditional training and real-world deployment. HuVAD prioritizes privacy by providing de-identified annotations and includes seven indoor/outdoor scenes, offering over 5x more pose-annotated frames than previous datasets. Our standard and continual benchmarks, utilize a comprehensive set of metrics, demonstrating that UCAL-enhanced models achieve superior performance in 82.14% of cases, setting a new state-of-the-art (SOTA). The dataset can be accessed at https://github.com/TeCSAR-UNCC/HuVAD.

Problem

• Identify abnormal human behaviors in videos
• Address challenges like diverse behaviors and rare anomalies
• Develop adaptive learning for real-world anomaly detection

Innovation

• Introduces HuVAD dataset with privacy-enhanced annotations
• Proposes UCAL framework for unsupervised continual anomaly learning
• Achieves superior performance in 82.14% of cases