OpenHSV

OpenHSV integrates high-speed videoendoscopy and synchronized audio capture with automated analysis to quantify vocal fold oscillations and laryngeal dynamics.

Key Features:

• State-of-the-Art Hardware Integration: Captures high-speed video and synchronized audio of vocal fold oscillations using commercially available equipment.
• Automated Data Analysis Pipeline: Executes a fully automatic processing workflow from data acquisition to derivation of quantitative laryngeal parameters.
• Advanced Video Segmentation: Applies efficient deep neural networks to segment the glottal area and produce glottal area waveforms and midline measurements.
• Quantitative Parameter Computation: Computes established quantitative parameters from video and audio data that are clinically relevant for assessing laryngeal function.

Scientific Applications:

• Voice Physiology Research: Enables detailed investigation of vocal fold oscillation mechanics and voice production using standardized high-speed videoendoscopy data.
• Clinical Diagnosis and Monitoring: Provides quantitative metrics to assist in diagnosing voice disorders and monitoring treatment outcomes at the laryngeal level.

Methodology:

Implemented in Python, OpenHSV uses a fully automatic analysis pipeline including video segmentation with efficient deep neural networks and computation of quantitative parameters from synchronized high-speed video and audio; performance was evaluated by analysis of image quality and derived quantitative parameters in a preliminary clinical study of 28 healthy subjects.