Automated Real-Time Tracking System for Blood Vessels of the Eye
MetadataShow full item record
The Eye tracking is a very hot topic and in the past few years, not-invasive methods have been introduced thank to the development of new image processing and computer vision techniques. Most of them are based on optical-tracking, which means that the tracking of the eye is performed by using infrared light and complex optics to measure the corneal reflection of the light, taking also into account the shifts of the center of the pupil over the time. A new not-invasive approach to this problem is described in this the sis, in which the purpose is to have a video with fixed images that allows the measurement of the conjunctiva blood vessel diameter and blood velocity. Therefore, the area of interest of the eye is the region of the blood vessels. The main idea of the new approach is to make the camera seeing always the same blood vessels area and performing a cross-correlation algorithm in order to track the movement of the patients eye. Depending on the system and the resources available, this algorithm could consider the whole image or a specific pattern such as the vessels intersection. The main goal is to reduce the usage of the registration algorithm as post-processing technique in order to build a video with fixed and stable images of the region of interest.