基于nnU-Net的乳腺DCE-MR图像中乳房和腺体自动分割

霍璐,胡晓欣,肖勤,顾雅佳,褚旭,姜娈

Automatic Segmentation of Breast and Fibroglandular Tissues in DCE-MR Images Based on nnU-Net

Lu HUO,Xiao-xin HU,Qin XIAO,Ya-jia GU,Xu CHU,Luan JIANG

表4 交叉验证中的分割性能评价（以DSC为例）

Table 4 Performance metrics for segmentation during cross validation (taking DSC as an example)

乳房分割 腺体分割 验证集 Fold 1 0.970±0.049 0.941±0.061 Fold 2 0.982±0.014 0.945±0.046 Fold 3 0.971±0.044 0.942±0.047 Fold 4 0.982±0.016 0.946±0.052 Fold 5 0.975±0.012 0.942±0.031 平均值±标准差 0.976±0.027 0.944±0.047 测试集 0.969±0.007 0.893±0.054