Fig. 8
From: A spine segmentation method based on scene aware fusion network
The mean DSC (%) for the most of individual vertebra segmentation visualization results. DSC is a commonly used indicator to evaluate the accuracy of segmentation, which can reflect the similarity between the model output and the ground-truth annotation results. SAFNet performs well in terms of vertebral segmentation visualization effect, can effectively extract and represent vertebral structure information, and achieves the highest average DSC for most single vertebral segmentation visualization results on multiple testing datasets, indicating that the model has high stability and generalization ability. (SAFNet denotes scene aware fusion network; DSC denotes Dice Similarity Coefficient.)