Combining PPE detection with gait and arm-movement analysis identifies workers when facial features are obscured in industrial video.
Image-derived hand pose is fused with self-tracking data to correct drift and preserve 6DoF interaction when tracking is blocked.
Recessed waterproof pool wall cameras combine AI image analysis and live streaming to detect drowning events and send alerts.
Image-based overlap analysis between the occupant upper body and seat belt improves proper-wear detection beyond buckle or weight sensors.
Body scanning and AI garment rendering improve virtual try-on accuracy by simulating fit, drape, movement, and wardrobe matching.
A ring-shaped display light with a dark center reduces optical noise, improving finger-reflected heart rate signal accuracy.
A unified codec model detects players and extracts foreground masks in one pass, cutting redundant sports video analysis for real-time use.
Real-time CCTV and robot boarding data let elevators skip full or unnecessary stops while improving passenger and robot space use.
Image-based AI checks whether workers near arc-flash electrical circuits wear PPE before permitting operation, reducing injury risk.
A shared attribute localization module uses spatial transformation and channel attention to improve pedestrian attribute accuracy with less computation.
Animal images with human-like features are excluded during training to reduce false human detection and improve classification accuracy.
Magnetic coupling and registration features enable tool-free sensor mounting on grid ceilings while preserving secure alignment and avoiding ceiling damage.
A face-covering transmitter chip pairs UID reading with biometric matching to keep digital wallet payments secure without mask removal.
Pixel-wise saliency maps let content services place secondary content in video frames without blocking important visual regions.
A shared virtual space lets viewers and other distributors commit virtual items across live streams, expanding support visibility without separate actions.
Thermal imaging and behavior analysis help detect livestock disease and reproductive events early while reducing manual monitoring in group stalls.
Continuous AI analysis of kitchen video and sensor data detects hygiene and safety risks in real time, reducing audit gaps and supervision burden.