Diseases, like cancer, can cause disruptive changes in our tissues and introduce anomalies. These anomalies can be hard to capture, but understanding them is essential for biologists to discover new therapies. Traditionally, tissue images are analyzed either through rule-based algorithms or supervised learning to provide new insights on tissue structure and their patterns of change. In this project, we propose a comprehensive, three-step unsupervised image processing pipeline to automate this process of tissue analysis and to help biologists better visualize the internal characteristics of tissue images.

Biology is everywhere, which makes biological research important. Through SuperUROP, I hope to gain more exposure in computer vision and learn how we can apply deep learning in applied biology. Specifically, I am excited to translate what I learned from class to actual research and work with scientists from the Broad Institute to contribute to the field of computational biology.