Fig. 1From: Precision of CT-derived alveolar recruitment assessed by human observers and a machine learning algorithm in moderate and severe ARDSLung segmentation masks performed by 3 different observers in 2 patients at 3 chest levels. Red regions are lung segmentation masks performed by human operator#1, blue ones are lung segmentation masks performed by human operator#2, and green ones are lung segmentation masks provided by the machine learning algorithm. The arrows denote obvious segmentation errors by the machine learning algorithmBack to article page