Fig. 1From: Learning rich features with hybrid loss for brain tumor segmentationExamples of the BRATS 2015 training set (best view in color), showing two HGG cases in the first two rows and two LGG cases in the last two rows. Each row represents (from left to right): T1c, T1, FLAIR, T2 and ground truth. Pixels labeled in black are background in the last column. Each of the other colors represents a tumor region: necrosis (blue), edema (green), non-enhancing (white) and enhancing (red)Back to article page