Congratulations to Elizabeth (Lizzie) Su for a successful UROP research experience and presentation! Lizzie used convolutional neural networks to predict injury prior to standard methods of detection. Results found that injury could be detected earlier than standard optical coherence tomography imaging. She is currently working on a variety of parameters to improve predictions with the CNN and the incorporation of transfer learning to manage small image datasets. She will continue the work and submit a publication this summer. Congratulations Lizzie!! You did an excellent job! Her poster can be viewed here.
Read more →
