






Automated analysis of pathology images enables quantification of image properties in a systematic way with the ability to learn new features to better distinguish different classes - even beyond the visual capability of human experts.
Machine learning can predict disease biomarkers to provide a better understanding of why drug efficacy varies from patient to patient. It can stratify patients according to risk and identify the best treatment for a particular patient.
Similar algorithms for quantifying images can also track the contributors to climate change. Remote sensing enables us to do this globally and may be one of the keys to mitigating or adapting to this global crisis.
Pathology and remote sensing images are large, multispectral, and heterogeneous. They are time-consuming to annotate and require domain knowledge to effectively model. While the terminology of each discipline is unique, the same algorithms can extract powerful insights to drive impact.