iMedic: Towards Smartphone-based Self-Auscultation Tool for AI-Powered Pediatric Respiratory Assessment

TL;DR

The paper presents iMedic, a smartphone-based self-auscultation system for AI-powered pediatric respiratory assessment that leverages domain generalization to fuse large electronic stethoscope data with smaller smartphone data. Using an Audio Spectrogram Transformer backbone and MixStyle, the approach learns robust cross-device features, enabling reliable abnormal-lung-sound detection from built-in smartphone microphones. The App guides users through symptom input and four-site lung sound recording, transmits data to an AI server for classification, and continuously improves with accumulating data. Experimental results show MixStyle-enhanced performance on smartphone recordings, though accuracy remains slightly below high-quality stethoscope models, and usability testing indicates favorable acceptance. The work highlights the potential of accessible smartphone-based diagnostics to reduce global health disparities in pediatric respiratory care.

Abstract

Respiratory auscultation is crucial for early detection of pediatric pneumonia, a condition that can quickly worsen without timely intervention. In areas with limited physician access, effective auscultation is challenging. We present a smartphone-based system that leverages built-in microphones and advanced deep learning algorithms to detect abnormal respiratory sounds indicative of pneumonia risk. Our end-to-end deep learning framework employs domain generalization to integrate a large electronic stethoscope dataset with a smaller smartphone-derived dataset, enabling robust feature learning for accurate respiratory assessments without expensive equipment. The accompanying mobile application guides caregivers in collecting high-quality lung sound samples and provides immediate feedback on potential pneumonia risks. User studies show strong classification performance and high acceptance, demonstrating the system's ability to facilitate proactive interventions and reduce preventable childhood pneumonia deaths. By seamlessly integrating into ubiquitous smartphones, this approach offers a promising avenue for more equitable and comprehensive remote pediatric care.

Figures (2)

• Figure 1: Overview of the deep learning model architecture, including the preprocessing, AST backbone, and MixStyle block.
• Figure 2: Screenshots of the application interface: (a) symptom selection screen, (b) recording site guidance screen, (c) real-time recording screen, (d) results display screen.