Development and testing of an open source mobile application for audiometry test result analysis and diagnosis...

Area of Science:

• Medical Technology
• Artificial Intelligence
• Ophthalmology

Background:

• Pure tone audiometry is the standard for assessing hearing impairments, with results visualized on audiograms.
• Existing methods for audiogram analysis can be time-consuming and prone to human error.

Purpose of the Study:

• To develop and validate an open-source Android mobile application for scanning and analyzing audiograms.
• To classify hearing loss types using artificial intelligence (AI) and image processing techniques.

Main Methods:

• The application utilizes YOLOv5, Optical Character Recognition (OCR), and Hough Transform for image processing and digitalization.
• A clinically validated AI model analyzes the digitized audiogram for hearing loss classification.
• Algorithms were optimized for mobile device functionality and tested on various smartphones.

Main Results:

• The mobile application successfully scans, digitizes, and classifies audiograms with consistent performance across different devices.
• The AI-driven analysis provides accurate classification of hearing loss types.
• The system demonstrated efficacy in assisting with audiometric test result interpretation.

Conclusions:

• The developed mobile application serves as an accessible AI-driven decision support system for hearing loss evaluation.
• Integration of this tool can reduce clinical workload, enhance diagnostic accuracy, and minimize errors in hearing assessments.
• This technology holds particular significance for improving audiological services in resource-limited settings.