Cancer Diagnostics Enabled by Machines: Finding and Classifying Single Tumor Cells in Human Tissue
, CIO, Mindpeak GmbH
Gain insight into how GPU computing and deep learning bring precision medicine into breast cancer diagnostics. Breast cancer diagnostics today relies on human medical experts who need to find and classify tumor cells in very large microscopical images. This manual analysis is not only challenging, tedious, and error-prone, but its precision is also limited by the small number of cells a human expert can process. Deep learning technology can perform the same task in the blink of an eye, enabling us to inspect more patient tissue and thereby increasing precision. We'll explain how we developed and validated a deep learning model that's used in clinical routine now. Learn about the different success factors for a deep learning project of this scale using GPU computing on terabytes of data. Effectively using NVIDIA docker images was key for us to achieve the necessary fast deep learning experimentation and development life cycle.