Voice pathology classification based on High-Speed Videoendoscopy

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Otolaryngology

Background:

• Vocal fold vibration impairments affect voice quality and require accurate diagnosis.
• Objective and automated methods for classifying these disorders are needed.

Purpose of the Study:

• To develop and validate an automated method for classifying healthy versus pathological vocal fold vibration.
• To assess the efficacy of using High-Speed Videoendoscopy (HSV) for objective voice disorder diagnosis.

Main Methods:

• Utilized High-Speed Videoendoscopy (HSV) of the larynx.
• Applied image segmentation techniques to HSV recordings.
• Extracted novel numerical parameters describing spatio-temporal vocal fold dynamics.
• Classified patients into healthy and pathological groups based on extracted parameters.

Main Results:

• Achieved a cross-validation classification accuracy of 73.3%.
• Demonstrated the ability of novel numerical parameters to differentiate between normal and pathological vocal fold vibrations.

Conclusions:

• The proposed method offers a promising approach for the automatic diagnosis of voice disorders.
• Automated analysis of HSV data can aid in objective clinical assessment of laryngeal function.