Smart Endotracheal Intubation: Real-Time Visual Feedback and Edge Detection for Enhanced Procedure Guidance

Endotracheal intubation (ETI) is a medical procedure that has gained researchers' interest in smart healthcare. Using endoscopy powered with Artificial Intelligence (AI) technology to detect anatomical features, making ETI smart is possible. However, the unavailability of public datasets and the insufficiency of predictive AI models for medical applications make this a challenging problem. This paper proposes a proof of concept powered by an ODROID XU4 microprocessor to provide visual feedback to the physician during the intubation procedure. The visual feedback is processed in real time using an edge detection technique, which assists the physician in visualizing the patient's vocal cords. The platform is capable of hosting deep learning models that utilize visual feedback and the detected edges to highlight anatomical features, and it is tested on a realistic manikin of a human airway.