Advanced computing solutions for analysis of laryngeal disorders.

Area of Science:

• Medical Imaging and Computational Pathology
• Otolaryngology and Speech Science

Background:

• Clinical diagnosis of voice pathologies relies on analyzing laryngeal recordings (audio, color, shape, vibration) from medical imaging devices.
• Existing computational solutions for laryngeal disorder analysis face challenges and open issues.

Purpose of the Study:

• To examine state-of-the-art computational methods for laryngeal disorder analysis.
• To propose a categorical representation of significant published applications.
• To reveal opportunities for vision-based computerized solutions in laryngeal disorder evaluation and prevention.

Main Methods:

• Categorical representation of published applications based on scope, methodology, and results.
• Review of laryngeal image/video analysis in four categories: vocal fold segmentation, disorder classification, vibration analysis, and image stitching.

Main Results:

• Identified key applications and their characteristics in laryngeal image/video analysis.
• Highlighted limitations and areas for improvement in current computational approaches.
• Demonstrated the potential of vision-based computerized solutions.

Conclusions:

• Vision-based computerized solutions offer significant potential for the evaluation, early diagnosis, and prevention of laryngeal disorders.
• Further research in areas like segmentation, classification, vibration analysis, and image stitching can advance diagnostic capabilities.