Effect of Functional Endoscopic Sinus Surgery on Voice and Speech Recognition

Area of Science:

• Otolaryngology
• Speech Science
• Biometrics

Background:

• Chronic rhinosinusitis and nasal polyposis are commonly treated with Functional Endoscopic Sinus Surgery (FESS).
• Voice and speech alterations can occur post-FESS, potentially impacting speaker identification systems.

Purpose of the Study:

• To evaluate the influence of FESS on voice acoustic parameters.
• To assess the implications of these vocal changes on speaker identification and verification systems.

Main Methods:

• Prospective study comparing FESS patients with a control group.
• Voice recordings analyzed for acoustic parameters (e.g., fundamental frequency) and GRBAS scale.
• Speaker identification accuracy measured using machine learning and equal error rate (EER).
• Data collected pre-surgery, 2 weeks post-surgery, and 3 months post-surgery.

Main Results:

• Significant changes in Grade, Roughness, Breathiness, Asthenia, and Strain (GRBAS) observed post-FESS.
• A significant decrease in fundamental frequency (F0) was noted in the FESS group compared to controls.
• The equal error rate (EER) for speaker identification was higher in patients who underwent FESS.

Conclusions:

• FESS leads to a decrease in fundamental frequency (F0) and alterations in the vocal tract.
• These vocal changes result in increased error rates for speaker recognition systems in FESS patients.