From Offline to Periodic Adaptation for Pose-Based Shoplifting Detection in Real-world Retail Security

Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• The proposed periodic adaptation framework consistently outperformed offline baselines on AUC-ROC and AUC-PR in 91.6% of evaluations for pose-based shoplifting detection.result 95%
• The periodic adaptation framework enables edge devices to adapt from streaming, unlabeled data for scalable and low-latency anomaly detection.method 90%
• Each training update for the periodic adaptation framework completes in under 30 minutes on edge-grade hardware.technical 95%
• The paper introduces RetailS, a new large-scale real-world shoplifting dataset collected under multi-day, multi-camera conditions.method 90%
• The solution is designed for on-site Internet of Things (IoT) deployment in smart retail environments.technical 90%
• The proposed shoplifting detection approach is cast as a pose-based, unsupervised video anomaly detection problem.method 90%
• The solution may struggle in environments with poor lighting or unusual layouts.limitation 80%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card