AI Powered Wireless Stethoscope for Real-Time Asthma Detection

View at publisher Summarise with VIRAT
Authors: M. Ravindran, Dhilakshan Rasa Kumara, Rickesh Krishnarajah, Sujani Gunaseelan, Nithusiga Sivapaskaran, Mukunthan Tharmakulasingam
Cited by: 0
References: 19

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This paper presents the development of a cost-effective, wireless AI-powered stethoscope specifically designed for binary classification of respiratory sounds to detect asthma through advanced sound capture and deep learning algorithms. The device employs an INMP441 MEMS microphone for high-fidelity acoustic signal capture, an ESP32-WROOM-32U micro-controller for processing and Wi-Fi transmission, and a mobile application integrated with AI models capable of performing real-time binary classification to distinguish between asthmatic and non-asthmatic conditions. The system includes a Battery Management System for safe power operation and a custom Printed Circuit Board for compact integration. Four different AI architectures were evaluated: LSTM-based models with YAMNet embeddings, 2D Spectrogram-based CNN, VGG16-based models, and Hybrid 1D CNN-LSTM approaches. The 2D Spectrogram-based Convolutional Neural Network achieved the highest accuracy of 89% with precision of 89%, recall of 88%, F1-score of 88%, ROC-AUC of 0.95, PR-AUC of 0.954, and a Brier Score of 0.0843. The system exhibits a wireless transmission latency of 215ms and a significantly lower latency of 30ms in wired mode. This portable diagnostic tool addresses the high cost barriers of existing digital stethoscopes while enhancing healthcare accessibility in remote and underserved regions.

Fields of Study & Subjects

Phonocardiography and Auscultation Techniques Artificial Intelligence in Healthcare ECG Monitoring and Analysis

Publication details

Journal / source
IEEE International Conference on Industrial and Information Systems (ICIIS)
Pages
43–48
Publication date
16 January 2026
Language
EN