Tearney Lab, Harvard Medical School

Overview

Developed a deep learning pipeline for Parkinson's disease diagnosis via esophageal OCT image segmentation. Presented research at the Wellman Center for Photomedicine at MGH.

Contributions

• Built deep learning pipeline for Parkinson's diagnosis via esophageal OCT segmentation
• Translated 16+ clinician requirements into model architecture and training strategy
• Achieved 85%+ classification accuracy with LT-U-Net across all classes
• Presented at the Wellman Center for Photomedicine, Massachusetts General Hospital

Tools and methods: Python, PyTorch, OCT Imaging, LT-U-Net, Deep Learning, Medical Imaging.