Used classification and regression methods to predict valence after screening the data with ANOVA in R. With classification methods, multilayer neural network and SVM were used, and prediction results were above the chance level. With regression methods, linear regression was also used and similar prediction result were generated.

Retina Scan Image Analysis to Detect Presence of Glaucoma and Measure Length of Blood Vessels.

This project focuses on developing a software to detect whether there is glaucoma and measure vessel length in a retinal scan image in Python. Used Convolutional Neural Network to develop two models to detect glaucoma based on the radius of optic disc and cup, and the prediction accuracy is relatively high (around 0.8-0.9). Blood vessels length were measured using segmentation methods from OpenCV package. A GUI was also developed using Python.