Fig. 2.
Examples of manual and automatic segmentations on the testing sets from dataset1 and dataset2. Illustrative subjects with the highest, median, and lowest Dice are presented from top to bottom in each dataset box. Each line corresponds to one subject. The manual segmentation of annotators 1 and 2 are respectively in green and yellow. The prediction of the 2-step method trained with data augmentation is presented in red, thresholded at 0.5, and denoted as “prediction”. The reported Dice is between the prediction and the manual segmentation. When there are two annotators, the Dice shown is computed between the prediction and the 1st annotator. (2 column fitting image). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)