Project · research
Renal Cell Carcinoma Classification
A medical imaging pipeline using transfer learning to classify renal cell carcinoma grade.
Project overview
This Informatics Skunkworks project, completed with the UW–Madison School of Medicine, explored automated grading of renal cell carcinoma from CT imagery.
Technical work
I optimized a segmentation script to isolate kidney regions and built a pipeline that converted three-dimensional MATLAB volumes into selected two-dimensional slices. We then adapted a pretrained ResNet to the grayscale and textural characteristics of medical imagery.
Outcome
The resulting workflow connected raw volumetric medical data to a standard transfer-learning architecture and established a foundation for non-invasive radiomic classification.