Innovation
Exploring hybrid methodologies for advanced surgical data analysis.
"Federated Learning-Based Collaborative Alert System for Emergency Medical Devices" (IEEE IoT Journal 2024):
Resolves data silos across hospitals via lightweight CNN-LSTM models predicting defibrillator battery failures (92.3% accuracy). Techniques are transferable to privacy-preserving computing in OR IoT systems.
"RFID-Computer Vision Fusion for Surgical Instrument Tracking" (Nature Communications 2023):
Proposes a dual-frequency RFID/YOLOv7 calibration algorithm, reducing instrument localization error from ±15cm to ±2cm. Provides methodological foundations for multi-source data fusion.
"Quantifying Ethical Decision Boundaries for Medical Large Models" (ACM FAccT 2024):
Models AI acceptance rates versus surgeon experience using Bayesian networks, finding 40% higher compliance challenges from senior doctors. Supports responsibility boundary design in this project.