Sleep apnea is a discret disease, many people are affected by this disease but few of them are getting diagnosed by practicians. One of the reasons that could explain that is the lack of information on the subject but also the complexity to diagnose the pathology that need a certain expertise from the physician. In order to be able to interpret the different tests done by the patient, the practician need to have an experience to interpret the patient's curve. That's where AI come to help, With a good classification done by several expert of the domain, it could be possible to create a model that will identify the region where there is a patern of sleep apnea in the curve, therefore the test interpretation could be done in an easier way in order to facilitate this access to healtcare.
Now that we have understand what is the pathology, i will explain how to diagnose it. the patient must do some tests such as the polysomnography that will record different informations about the patient, how he breath, does the patient move during his sleep etc... different criterias gives the sign of respiratory problems when the physician monitor the curves obtained by the polysomnography.
As you can see in the video, we can find pattern to recognize the pathology, but this need a lot of experience and it could be done using AI that could use a dataset of curves with sleep apnea and curves of normal sleep, alowing to analyse the curve faster and letting the physician focus on the most added value job, the patient care. Hence we could use AI in order to classify the important moment of the polysomnography and maybe to find new paterns with a neural networks that could make a correlation between new criteria to understand better the pathology
In order to develop our tool, we will use different javascript libraries and models avaibles online such as :
