Fig. 1From: Utility of a novel integrated deep convolutional neural network for the segmentation of hip joint from computed tomography images in the preoperative planning of total hip arthroplastyThe flowchart of our proposed segmentation method. The input CT to the upward network has window level 40, window width 200 to highlight all bone structure. And the CT input to femur segmentation network has window level 500, window width 1000 to remove soft tissues and preserve bone structure as much as possibleBack to article page