Pathologists assess tissue samples on glass slides via microscope, but this process suffers from intra-/inter-observer variability. We develop deep learning-based tissue segmentation models to objectively and reproducibly quantify histopathology images from various cancer types.