Vocal Cords Detection Using Edge Detection and Machine Learning Filtration

Automating endotracheal intubation (EI), a critical medical procedure required for lung ventilation, has gained significant attention in the realm of medical imaging and robotics. Central to the success of automated EI is the reliable detection and visualization of vocal cords, a step traditionally performed using laryngoscopy. This paper introduces a novel method for detecting and segmenting vocal cords in endoscopic images by leveraging advanced edge detection techniques combined with machine learning (ML) filters. This approach uniquely integrates geometric and textural features derived from contours identified in the image, distinguishing vocal cords with high precision and speed. Beyond automating EI, this methodology has broader applications, including the removal of foreign objects and facilitating tracheal endoscopy. This paper details the development of a specialized dataset and the implementation of the solution, emphasizing its algorithmic innovations. The results demonstrate the efficacy of this approach in improving the accuracy and efficiency of vocal cords detection, offering a repeatable strategy for similar medical imaging challenges.