Fig. 4From: 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 arthroplastyCMG network ensures the overall accuracy of segmented femur head. A Qualitative comparison of the segmentation results on a given axial slice of the diseased hip joints (ONFH necrosis of the femoral head, FNF femoral neck fracture, DDH development dysplasia hip, OA hip osteoarthritis). B Accuracy (DOC, %) comparison between the proposed method and four state-of-the-art methods on diseased hip joints. C Accuracy (ASD, px) comparison between the proposed method and four state-of-the-art methods on diseased hip joints. D Accuracy (HD, px) comparison between the proposed method and four state-of-the-art methods on diseased hip joints. **p < 0.01, ***p < 0.001, ****p < 0.0001 versus CMG NetBack to article page