Naimul Mefraz Khan
Software Systems
monitoring vital signs such as heartrate, blood pressure, oxygen saturation is a very important aspect in different application areas such as remote patient monitoring, exercise, baby monitoring etc. Traditional means of vital signs monitoring requires attaching a sensor to the subject under monitor. An alternative is an entirely computer vision-based approach, where the vital signs are extracted just from an image/video.
To create a vital signs detection app (mobile or web) that can detect and display multiple vital signs in real-time through a video stream from a mobile or web camera.
1. A mobile or web camera streams video to the app.
2. The app utilizes machine learning algorithms to extract vital signs from the image/video.
3. The extracted vital signs are shown as an overlay on the video stream.
4. Any urgent issues detected in the vital signs (e.g. high heartrate) is displayed with an alert.
1. Study vital signs detection algorithms. Many existing ML approaches can do this from camera images.
2. Find an algorithm that is fast enough to do this in real-time.
3. Validate the approach with benchmark datasets.
4. Create a demonstrable product where the extracted vital signs can be validated by simultaneously wearing a wearable such as a smartwatch.
See Suggested Approach
Study existing algorithms to decide on the fastest approach
Study existing algorithms to decide on the fastest approach
Displaying vital signs on the user interface, create alerts for critical events
Web/mobile app development
NMK08: Vital signs detection through a camera | Naimul Mefraz Khan | Monday August 30th 2021 at 12:43 PM