)
)
Mobile-accessible deep learning-based self-assessment tool for measles screening in low-resource settings
-
Measles is highly contagious and poses serious health risks, particularly to young children. Global vaccination coverage declined in 2023, with only 83% receiving the first dose and 74% the second, especially in low-income countries. Traditional diagnostic tools are often unavailable in low-resource settings. While deep learning has been used to detect measles skin lesions, prior studies relied on small datasets and lacked validation across key demographic groups.
-
This study developed and validated a deep learning model based on over 44 000 skin lesion images, demonstrating robust performance across age, gender, origin, skin tone, body region and rash colour. The model was integrated into a rule-based, mobile-accessible self-assessment tool that combines image analysis with a structured questionnaire on symptoms and exposure—designed for practical use in low-resource settings.
-
The tool provides a scalable, low-cost solution for early measles screening and supports timely detection and outbreak response in low-resource settings. It is suitable for both the general public and primary healthcare workers and could inform future digital health strategies in infectious disease surveillance and triage.