An ordinary smartphone can help parents notice breathing problems in their child in time

Edited by: Tetiana Pin

Scientists from Seoul National University of Science and Technology, led by Associate Professor Kim Seong-eun, have created a system that uses recordings from an ordinary phone microphone to recognize alarming sounds in the lungs of children aged 0 to 6 years.

The main difficulty was that neural networks are usually trained on recordings from electronic stethoscopes. But a phone hears sound quite differently — it has a different microphone and different settings. Therefore, a model that works well with a stethoscope begins to make mistakes on recordings from a phone.

To fix this, the researchers took two steps. First, they processed the phone recordings: they removed excess high frequencies and reduced the sound quality so that it became closer to that of a stethoscope. Then they added a special module that helps the model better understand "phone" sound. This refinement was built into a system that was already able to analyze sounds.

For training, they collected data from 611 children. For 147 of them, recordings were made with both a stethoscope and a phone at the same points. Separately, they tested the system on 48 children who were recorded only with a phone in another location.

The result was good. On the main phone recordings, the system correctly recognized problems with an accuracy of about 80%. The model that was trained only on phone recordings performed worse. And if you simply mix stethoscope and phone data without special processing, the result was even weaker.

On new, "unfamiliar" recordings, the system also maintained normal accuracy. But a model trained only on a stethoscope almost completely stopped working. Without special adaptation, moving from one device to another is almost impossible.

The authors emphasize: this is not a replacement for a doctor, but an assistant. The system can suggest that something has changed between visits, but the final decision is always made by the pediatrician.

Why does this matter? A stethoscope is the main way to check the lungs, but for that you need to go to a doctor. Electronic stethoscopes are expensive, while almost everyone has a smartphone. Early detection of breathing problems in young children is especially important because the condition can deteriorate quickly.

This work continues previous research by the same group. Earlier, they had already shown a prototype of an app that records sounds and sends them to a server for analysis. Now the system copes better with different recording conditions.

For parents, the main takeaway is simple: such technology is a reason to watch the child more closely, not a reason to postpone a visit to the doctor. If a child is coughing or having trouble breathing, the decision always remains with the specialist.

5 Views

Sources

  • AI turns smartphones into pediatric lung sound screeners - ModeZone

  • Smartphone-Based Detection of Abnormal Lung Sounds Using Domain-Adaptive Deep Learning (DOI 10.1109/JBHI.2026.3718492)

  • PubMed: Smartphone-Based Detection of Abnormal Lung Sounds Using Domain-Adaptive Deep Learning

  • Adapted AI improves smartphone detection of abnormal lung sounds in children (Medical Xpress)

  • Adapted AI Improves Smartphone Detection of Abnormal Pediatric Lung Sounds (hyper.ai)

  • AI Tool Improves Smartphone Detection of Abnormal Pediatric Lung Sounds (Hospimedica)

  • Could a smartphone listen for childhood lung problems? AI narrows the gap (Digital Journal)

  • iMedic: Towards Smartphone-based Self-Auscultation Tool for AI-Powered Pediatric Respiratory Assessment (arXiv)

  • iMedic (Seoul National University of Science and Technology repository)

Comments

Read more articles on this topic:

Did you find an error or inaccuracy?We will consider your comments as soon as possible.