Researchers have developed a novel two-stage pipeline for automated glottal area segmentation from high-speed videoendoscopy. This system, which combines a YOLOv8n localizer with a U-Net segmenter, achieved high accuracy on established datasets, with Dice Similarity Coefficients reaching up to 0.856. An initial clinical study indicated that the glottal area Coefficient of Variation could effectively distinguish between healthy and pathological vocal functions. The pipeline operates at approximately 35 frames per second on standard hardware, facilitating real-time clinical review and consistent extraction of laryngeal kinematic measures. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT This new segmentation pipeline could improve the accuracy and efficiency of diagnosing laryngeal pathologies.
RANK_REASON This is a research paper detailing a new methodology and its evaluation. [lever_c_demoted from research: ic=1 ai=1.0]