Digital Stethoscope: from phone + headset + ear horn

Digital Stethoscope: from phone + headset + ear horn

A cheap omnipresent digital stethoscope could help to fight COVID-19 as well as other lungs diseases. There are some products, but they are not cheap. There are some DIY projects, but they are not active. Here I am trying to initiate a feasibility talk because some ML applications could be for good use.

       How will it help:
    • Many cases are mild or asymptomatic. In Australia and UK, they test only if you have a fever, contacted or was abroad (so far). A mild suspect just self isolates for 2 weeks (or less) and then free to go. However, there are family members, and both parties may be loose on proper internal isolation. Even 70-80% as precise, but objective quick test could change behaviour or be used in pre-screening.
    • Routing monitoring of active cases: patients can make that routine themselves to save time for a doctor, who just see a case progress dashboard. No extra CT or other lugs scans.
    • In some cases, virus progress rapidly, an easy check may help to decide to visit a hospital
    • The cheaper the device the easier to spread monitoring in developing countries. Pass graph results / sound recordings or touch to hear in online video GP assessment. Many other illness diagnostics could get huge boots if such a device is as cheap and present as a body thermometer.

       The existing devices: 

They are expensive for people to buy, just to test. However, we may pursue those who have mild or other cases to monitor themselves and get decent data. ( gather that data with, but the android version is needed). The range is mostly above $300, with only one for $60. If you know more, especially cheap ready to use, please, share. They have an android app.


COVID usage example:




VitalSigns-Stethoscope. The product No longer available ~?

          The DIY projects:



iStethoscope Pro – low rated app

      My thoughts:

Disclaimer: I didn’t spend much time on research and properly study mentioned above DIY projects.
My own attempts to make recordings with some plastic flute, phone mic, headset, cone from an aluminium can FAIL to produce decent heart-lungs sounds. It could be that just a headset mic is not sensitive enough, or different cone shape / material is needed. At the same time doctors diagnose just with their own ears, so such a mass test device I believe could be made. Even if we can get enough data and show a pre-screen COVID diagnose models (from professionally made devices), it will spoor china manufacturers to make digital stethoscopes accessible for future waves.

Please, feel free to elaborate, pass and promote the idea to universities, maker clubs or critic.

         Additional information

ANZICS guidelines recommend minimization of stethoscope usage. This is actually another reason to have a personal one: no contamination but more data, better ML diagnosis.

Sound data. (there are some android apps with breathing examples)

Just a physical one, but could be useful to produce recordings.

        Similar idea - COVID by voice

Detecting COVID by voice coughs. They are currently gathering data, so feel free to help. I think a Stethoscope should better diagnose than just voice, besides it is still useful for other illnesses.
COVID-19 Sounds App. Cambridge.

EFPL A new app can help detect the coronavirus

COVID Voice Detector

        Spirometer test. 

How hard and how quickly you can expel air from your lungs. There are also some devices on the market, but I think some whistle like a cheap tube or visual vibrations to create the test curve proxy (the phone app, analysing the sound – vibrations).

         Ultrasound project

1 Like

A few links in addition:

Good ML example: noise and hart beat cancelling. Great work Johns Hopkins University!:

Another product with AI (trial faze). Claim 29% better than regular ears.

My intuition a stetho-vest with 3d kind of result should have even better ML diagnostics. I mean, several digital stethoscopes around the chest record breathing simultaneously.

MIT in 2015 started a good low-cost solution, but don’t see the current outcome.

LA, USA ~start-up

Existing products:

This digital version too cheap to be true ( $0.1), but could be still an easy option to digitize the data.

Some other $50-100 and more expensive digital stethoscopes.

A proper mic with an analog stethoscope membrane looks to be the easiest DIY version: