The purpose of this project is to monitor the eyelid movement in order to detect diseases or medical conditions.
The detection is made with the EMM - the Eyelid Motion Monitor device, that was designed and build in the High Speed Digital Systems laboratory in previous projects.
Apparently, many conditions can effect the eyelid movements, such as: damage to muscles or nerves, eye disease (like ptosis and cataract), systemic disease (like grave’s disease) and neurological disease (like Parkinson), and even medications can affect our blinks.
The problem is that all of those diagnosis are observation based, which is an inaccurate method and not a quick one.
All of these lead to the idea of the EMM – the Eyelid Motion Monitor.
A magnet is glued to each of the patient’s eyelids. That way, movements of the eyelid changes the magnetic field around the eye.
Beside the magnet, the patient wears glasses containing sensors on them (4 sensors around each eye). The sensors measure the magnetic field created and a digital card samples it.
All the data can be saved on an SD card for later use or transferred in a real time.
The algorithm that have been made in this project is getting the data from the SD card.
It detects the blinks of each measurement, and extract 8 parameters from each blink (amplitude, energy, opening and closing duration and speed etc.).
From comparing each pair of parameters the algorithm gets ellipses parameters that are derived by distribution’s characteristics of the measurement.
From all of those ellipses, the vision of this project is to map the parameters space to areas. Each area indicate medical condition.
Once we have a measurement we can classify to which area it belongs to.
The second part of the project examine the main assumption in the algorithm, normal distribution of the data.
Furthermore, we were taking advantage of all eight parameters by combining the parameter space with PCA and apply the ellipse model in the first two dimension of PCA space.
