OpenHSV: an open platform for laryngeal high-speed videoendoscopy

Area of Science:

• Laryngology
• Voice Science
• Biomedical Engineering

Background:

• High-speed videoendoscopy (HSV) is crucial for studying laryngeal dynamics and voice disorders.
• Existing HSV research tools lack open-source accessibility, expandability, and integrated data analysis.
• There is a need for standardized, advanced platforms for HSV data acquisition and analysis.

Purpose of the Study:

• To introduce OpenHSV, an open-source research platform for high-speed videoendoscopy.
• To provide a fully automatic data analysis pipeline for laryngeal dynamics.
• To establish a standardized tool for voice scientists and clinicians using HSV.

Main Methods:

• Developed OpenHSV using state-of-the-art commercial hardware and a Python-based graphical user interface.
• Implemented synchronized recording of high-speed video and audio data.
• Utilized deep neural networks for glottal area segmentation and derived quantitative parameters.

Main Results:

• OpenHSV successfully acquired and analyzed synchronized video and audio data from 28 healthy subjects.
• The platform demonstrated applicability, performance, and usefulness through image quality and parameter analysis.
• Quantitative, clinically relevant video and audio parameters were computed automatically.

Conclusions:

• OpenHSV offers a valid and standardized approach to high-speed videoendoscopy data acquisition and analysis.
• The platform serves as a valuable research tool for understanding voice physiology and diagnosing voice impairments.
• OpenHSV is poised to become the foundation for future clinical HSV systems.