A Survey on Machine Learning Approaches for Automatic Detection of Voice Disorders.

Area of Science:

• Biomedical Engineering
• Computational Linguistics
• Speech Science

Background:

• The human voice production system is complex and susceptible to damage affecting vocal folds.
• Voice changes impact physical functioning and emotional well-being, necessitating early detection.
• Timely identification of voice disorders is vital for patient recovery and quality of life.

Purpose of the Study:

• To survey research on automatic detection of voice disorders using machine learning.
• To explore the capabilities of machine learning in identifying various voice disorder types.
• To analyze databases, feature extraction, and machine learning approaches in this field.

Main Methods:

• Systematic review of existing research on automated voice disorder detection.
• Analysis of machine learning algorithms applied to voice disorder classification.
• Examination of commonly used datasets and feature extraction methods.

Main Results:

• Machine learning techniques show significant promise in aiding the early diagnosis of voice disorders.
• Various approaches have been investigated for identifying different types of voice disorders.
• The study provides an overview of the current landscape of automated voice disorder detection.

Conclusions:

• Automatic detection systems can significantly ease the diagnostic process for clinicians.
• Further research and standardization of methods are needed to enhance accuracy and reliability.
• Machine learning offers a powerful tool for improving patient outcomes through early intervention.