Early Prediction of CVD Using Retinal Vessel Features
It has been shown that the retinal vasculature reflects the health of the microvasculature of the brain, heart, and other organs, thus providing a non-invasive means for evaluating the status of the cardiovascular system. Our goal is to use digital retinal imaging and automated retinal vascular network analysis techniques for screening early sign of CVD in retina. The computer-aided screening of systemic CVD is based on automatic segmentation of retinal vasculature and quantitative analysis of the global branching network pattern and local vessel structure features.
![]()

Cardiovascular disease (CVD) is the leading cause of death in the U.S., according to the Centers for Disease Control (CDC). In 2007, more than one in four deaths in the U.S. was attributable to heart disease. Our system has the potential to give patients an opportunity to make diet or life style changes and reduce the huge cost burden of CVD. The system will be an invaluable tool for CVD prevention and health maintenance.
Carotid Artery Plaque Tracking
The objective of our research is to identify plaque surface motion patterns that can be used to differentiate between symptomatic and asymptomatic cases, and quantify increases in transient ischemic attack (TIA) or stroke risk. Moreover, the quantification of elevated risk will allow the treatment of patients before they develop symptoms saving many lives since the mortality of stroke is extremely high with most of the patients suffering a stroke from previously asymptomatic lesions. Our methodology presents a new high-resolution, low-cost, non-invasive approach to identifying vulnerable plaques that lead to stroke. Using optical flow techniques and a novel Amplitude Modulation Frequency Modulation (AM-FM) method for motion estimation, the system will provide reliable motion vectors of ultrasound videos of carotid artery (CA) plaques. Our methodology can measure plaque motion and deformation at the pixel level by providing the plaque principal axis of deformation and producing elastography maps for each time step or frame of the B-mode US sequence calculating detailed maps of how plaque motion and deformation change throughout the cardiac cycle.


Preliminary strain results on asymptomatic plaque: (L) Principal axes of deformation of the plaque. Vector magnitude is logarithmically transformed for visualization. Regions of similar deformation are marked. C and D show the greatest strain due to motion. (R) Strain matrix trace plot indicating regions of strong relative area deformations.
