Skip to main content
. 2023 Apr 10;13(4):976. doi: 10.3390/life13040976

Table 2.

Evaluation of different segmentation methods for Dataset A.

% BG NFL GCL+IPL INL OPL ONL+IS OS OPR RPE
FCN Precision 99.73 80.04 97.51 75.87 69.13 98.24 75.88 87.19 75.61
Recall 97.09 93.52 90.12 93.15 80.97 92.86 86.76 81.59 92.78
DSC 98.42 86.26 94.13 83.63 74.58 95.48 80.96 84.30 83.32
Unet Precision 99.82 81.89 98.11 84.61 72.43 98.30 78.40 86.26 79.52
Recall 97.47 95.37 93.64 88.87 85.44 94.04 84.73 87.44 93.25
DSC 98.63 88.12 95.83 86.69 78.82 96.13 81.44 86.85 85.84
SegNet Precision 99.80 80.17 97.25 87.83 69.60 96.96 78.49 81.94 81.14
Recall 97.48 95.93 93.89 85.35 83.67 93.26 81.76 87.04 91.91
DSC 98.63 87.76 96.04 86.57 76.80 95.08 81.12 84.41 86.90
TransUnet Precision 99.42 81.18 97.38 88.07 67.53 97.00 73.25 84.71 78.58
Recall 97.08 95.75 93.34 82.69 85.09 93.76 88.66 88.12 89.98
DSC 98.23 88.45 95.32 85.29 75.30 95.35 80.22 86.38 83.90
TranSegNet Precision 99.75 82.18 97.91 89.29 77.40 98.21 77.98 91.30 80.46
Recall 97.99 96.36 94.33 91.29 85.37 94.56 89.83 85.32 95.56
DSC 98.86 87.76 96.64 90.28 80.27 96.35 82.14 88.21 87.36