Doctors that are performing research in Sudden Infant Death Syndrome (SIDS) need better tools for collecting and analysing data from infant sleep studies. Our client's theory is that a major factor in determining the risk of SIDS is the percent of time the infant spends in the Slow Wave Sleep (SWS) sleep state versus the Rapid Eye Movement (REM) sleep state and the sleeping pattern overall. The issue with analysing when the infant is in these sleep states is a challenging one. Parents do not want intrusive tools in their child’s sleep area and they want their child to be safe. Reliability and effectiveness of the monitoring product is essential also. Researchers are looking for a way to gather the sleep data of infants without inconveniencing the parents or child.
We created a system built upon previous work from another senior design team to record a sleeping infant and use the video to determine the percent of time the infant is in each sleep state. There is also charting of the detected activity level, heart rate and breathing rates that is displayed as a function of time.
This project consists of developing a web enabled tool by integrating a web camera with special software which graphically displays infant sleep data to researchers conducting Sudden Infant Death Syndrome (SIDS) Polysomnography (sleep studies). The project requires integration of existing open source components to detect heart rate and customizing a user interface to display the video, activity and heart rate graphs. This also includes developing an algorithm to combine the activity and heart rate data to determine sleep states as a function of time.
The team made use of special software developed by MIT called, Eulerian Video Magnification (EVM), which is available for non-commercial research purposes and can be downloaded from: http://people.csail.mit.edu/mrub/vidmag/#code. This software allowed for video enhancement, making motion and blood pressure (redness pulsing through visible skin) much more evident.
The system hardware components consists of a motion detection camera, remote server, user’s website and an optional wireless mobile device, to monitor infants as they sleep inside their cribs. Figure 1 in Appendix B shows the high-level system overview. The video data is currently uploaded to a server at SIDSKnowMore.net (hosted by 1and1.com). The researchers will use the system and data to evaluate the infant’s SWS/REM sleep states and cycles.