University of Windsor researchers released SH17, a 8,099-image open dataset for AI-driven detection of personal protective equipment (PPE) in manufacturing settings.
Key Details
- 1SH17 includes 8,099 images with 75,994 labeled PPE and body part instances across 17 categories.
- 2Dataset covers diverse PPE, such as helmets, gloves, safety glasses, and small items like earmuffs.
- 3AI models (e.g., YOLOv9) trained on SH17 achieved mean average precision of 70.9%.
- 4Images were sourced globally to reduce bias and reflect real-world manufacturing conditions.
- 5Both the dataset and trained model weights are publicly available for further research and development.
Why It Matters
SH17 offers a valuable resource for advancing vision-based PPE detection in industrial settings, demonstrating strong generalization and supporting scalable, non-invasive worker safety solutions. Its release enables further research and deployment of AI monitoring in manufacturing, potentially reducing accidents and improving compliance.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.