69 whole slide images for early gastric cancer diagnosis, evaluating the proposed variational energy network (VENet).
The self-collected 30000 pathological images for gland segmentation, including training images and annotations.