Vinitha Selvaraj

• Analyzed B-mode ultrasound images of carotid artery and implemented machine learning algorithms which utilize texture and intensity features to detect intraplaque hemorrhage and determine risk of stroke. • Classification accuracy of the algorithm was better than neurosonologists’ assessment on average. Recently, my work has been accepted to the International Stroke Conference 2019.

• Developed a tool to track the reflected Infrared LED glints on eyes in real time for the purpose of eye tracking. • Created and processed databases for calibration of the camera and testing. • Performed experiments to determine optimal camera position in headset, quantitative evaluation of gaze tracking algorithm on different users and blink detection.

• Worked on face detection and gender classification of real-time facial images. • Compared existing state-of-the-art machine learning techniques (PCA, LDA, SVM) and hybrid approaches to arrive at the best model.