Pranav Kolluri
Deep Learning to Accelerate Motion Grading of HR-pQCTs
High Resolution Peripheral Quantitative Computed Tomography, or HR-pQCT, is a medical imaging technique used to assess the architecture of the of the cortical and trabecular bone. Our lab, the Bone Quality Research Laboratory, uses this imaging technique in addition to MRIs to better understand how bone structure changes with disease. Typically, at the time of image capture, the operators have to manually (and subjectively) “score” if the scan has been impacted by motion artifacts. My work focuses on automating the scoring process of HR-pQCT capture via deep learning on a dataset of our own creation to remove this capture bottleneck for the lab and perhaps even more broadly.
Message To Sponsor
Thank you so much for your help! Having financial support means that I fully focus on my research and set myself up for the future! I've learned so much through URAP, both about my field as well as general research, and it's had a big impact on where I see myself in the next couple years!
Major: Electrical Engineering and Computer Science
Mentor: Galateia Kazakia, UCSF Radiology
Sponsor: Cheunkarndee Fund