cezius / Neonatal-Respiration-Monitoring-Algorithm

A real-time camera based respiration rate estimator algorithm for neonatal records

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Neonatal-Respiration-Monitoring-Algorithm

This is an implementation of the camera based Respiration Rate Estimator Algorithm which is described in the following publication: https://www.mdpi.com/2076-3417/11/16/7215 (within this, the algorithm is referred as Respiration Calculator Block).

The algorithm estimates the respiration rates from the respiration related motion that can be found on the ROI which is detected by a convolution neural network. This was designed to work on RGB frames that contain images from newborn babies in neonatal intensive care units. The input is a hdf5 file which contain the frames from the infant. The ROI-detector was trained to detect the belly or the back of the babies in daytime shots. The output are the estimated respiration rates i RPM (respiration per minute). For the details, see the article referred above!

Data

For privacy reasons, we are unable to share the data set currently.

How to run?

Step 1: Clone the repository!

Step 2: In the root of the cloned project, start a command line run the following:

jupyter notebook

Step 3: In the opened browser, open the following:

RunRespMonitoringOnMyHDF5.ipynb

About

A real-time camera based respiration rate estimator algorithm for neonatal records


Languages

Language:Jupyter Notebook 75.4%Language:Python 24.6%