Research Project Title:
Developing a Multipurpose Automated Platform for Quantitative Cellular Imaging
abstract:This projects seeks to develop deep learning models that accurately predict and classify immune cells. Existing solutions have limited breadth because they do not account for staining techniques and cell functions. We will build, train, and test a CNN model that identifies cellular and sub-cellular regions on diverse cell types and contexts. We will collect datasets of cellular movies and develop computer vision models that process and analyze such datasets. We will test this tool with specific case studies relevant to the field, and generate data visualizations that are user-friendly to researchers in a biology lab setting. This software platform will enable researchers, clinicians, and pathologists to conveniently understand disease-relevant cell properties from their own image sets.
Through the Super-UROP, I hope to learn how to perform image processing with Computer Vision, which I learned about from previous machine learning classes. This will also help me determine whether I’d like to pursue graduate school, and work on something I believe is impactful.