UMBC Computer Science Ph.D. student Adrian Rosebrock and the company he founded and heads, ShiftyBits LLC, were recently awarded a competitive research contract from the National Cancer Institutes (NCI/NIH) to conduct research and development in the use of image processing and machine learning techniques to automatically analyze histology images of the breast.
Through the awarded research contract, Adrian will be helping NCI researchers with the design and development of automatic identification techniques for terminal duct lobular units (TDLUs) of the breast, the structures from which most breast cancers arise. This work will also create standard metrics for TDLUs that will aide researchers working within cancer research.
Data suggests that the morphology of TLDUs is related to several breast cancer risk factors, including mammographic density. In addition, TDLU morphology may represent an independent risk factor for breast cancer among women with a biopsy for benign breast disease. For this research project, Shiftybits will be given access to a large NIH dataset of breast biopsies as well as the Komen histology datasets.
Adrian received a B.S. in Computer Science in 2010 from UMBC and founded Shiftbits, LLC in 2011. He is continuing his studies as a Computer Science Ph.D. student at UMBC student, focusing on the combination of text and image retrieval systems. One research project he at UMBC involves the automatic identification of pills in images. At UMBC, Adrian is and working with Professor Tim Oates and Dr. Jesus Caban who is a researcher at Naval Medical Center and NIH and also teaches at UMBC.