Table 6 Comparison of the proposed method with state-of-the-art methods in HOMOMO, Roboflow, and YouTube datasets.
Method | Datasets | |||
|---|---|---|---|---|
Criteria | HOMOMO | Roboflow | YouTube | |
Proposed | mIoU | 98.44% | 85.4% | 81.32% |
Mean accuracy | 99.5% | 94.6% | 93.0% | |
Mean precision | 98.98% | 86.98% | 85.73% | |
Mean recall | 98.52% | 86.76% | 84.76% | |
Mean specificity | 99.1% | 87.36% | 85.41% | |
Mean F-score | 98.74% | 86.86% | 85.24% | |
Mean F-boundary | 82.01% | 75.99% | 70.01% | |
Mean test time (s) | 0.011 | 0.018 | 0.012 | |
DeeplabV3(ResNet101)18 | mIoU | 84.23% | 68.5% | 61.9% |
Mean accuracy | 88.21% | 87.2% | 86.15% | |
Mean precision | 86.98% | 73.32% | 70.1% | |
Mean recall | 86.22% | 72.98% | 69% | |
Mean specificity | 86.83% | 73.60% | 70.08% | |
Mean F-score | 86.59% | 73.14% | 69.54% | |
Mean F-boundary | 71.23% | 53.54% | 51.02% | |
Mean test time (s) | 0.025 | 0.026 | 0.024 | |
U-Net18 | mIoU | 82.10% | 50.3% | 48.45% |
Mean accuracy | 86.90% | 83.4% | 82.71% | |
Mean precision | 86.20% | 67.13% | 59.98% | |
Mean recall | 84.99% | 65.34% | 58.18% | |
Mean specificity | 85.60% | 66.43% | 59.12% | |
Mean F-score | 85.59% | 66.22% | 59.06% | |
Mean F-boundary | 69.10% | 41.93% | 41.23% | |
Mean test time (s) | 0.043 | 0.039 | 0.040 | |
Mask2Former53 | mIoU | 87.11% | 69.14% | 70.43% |
Mean accuracy | 89.78% | 87.18% | 86.41% | |
Mean precision | 88.14% | 77.25% | 74.01% | |
Mean recall | 87.59% | 75.12% | 73.16% | |
Mean specificity | 88.15% | 76.71% | 74.31% | |
Mean F-score | 87.84% | 76.17% | 73.58% | |
Mean F-boundary | 70.2% | 56.18% | 57.31% | |
Mean test time (s) | 0.030 | 0.031 | 0.029 | |
SwinUNet54 | mIoU | 90.01% | 72.19% | 71.31% |
Mean accuracy | 91.13% | 88.34% | 87.12% | |
Mean precision | 90.39% | 74.65% | 73.03% | |
Mean recall | 90.23% | 72.49% | 72.12% | |
Mean specificity | 90.96% | 73.11% | 73.07% | |
Mean F-score | 90.31% | 73.55% | 72.57% | |
Mean F-boundary | 74.41% | 58.98% | 60.71% | |
Mean test time (s) | 0.033 | 0.032 | 0.31 | |
TransUNet55 | mIoU | 88.78% | 71.43% | 72.1% |
Mean accuracy | 88.99% | 88.98% | 89.11% | |
Mean precision | 88.91% | 73.67% | 73.26% | |
Mean recall | 88.86% | 72.98% | 73.19% | |
Mean specificity | 88.93% | 73.14% | 74.38% | |
Mean F-score | 88.88% | 73.32% | 73.22% | |
Mean F-boundary | 76.67% | 60.34% | 59.98% | |
Mean test time (s) | 0.036 | 0.0321 | 0.032 | |
YOLOv519 | mIoU | 80.02% | 61.27% | 60.94% |
Mean accuracy | 87.02% | 80.63% | 83.45% | |
Mean precision | 86.14% | 70.34% | 69.91% | |
Mean recall | 84.31% | 69.76% | 69.32% | |
Mean specificity | 84.90% | 70.58% | 70.1% | |
Mean F-score | 85.21% | 70.04% | 69.61% | |
Mean F-boundary | 63.12% | 54.11% | 58.43% | |
Mean test time (s) | 0.004 | 0.006 | 0.006 | |
YOLOv1156 | mIoU | 82.96% | 62.78% | 63.01% |
Mean accuracy | 89.62% | 83.03% | 85.14% | |
Mean precision | 89.14% | 72.78% | 72.44% | |
Mean recall | 87.21% | 71.44% | 70.99% | |
Mean specificity | 87.85% | 72.23% | 71.97% | |
Mean F-score | 88.16% | 72.10% | 71.70% | |
Mean F-boundary | 68.13% | 55.91% | 57.87% | |
Mean test time (s) | 0.0029 | 0.0025 | 0.0029 | |
YOLOv1256 | mIoU | 81.12% | 63.01% | 62.11% |
Mean accuracy | 89.43% | 83.13% | 85.41% | |
Mean precision | 88.14% | 72.65% | 73.06% | |
Mean recall | 79.87% | 68.03% | 68.31% | |
Mean specificity | 80.12% | 69.17% | 69.48% | |
Mean F-score | 83.80% | 70.46% | 70.60% | |
Mean F-boundary | 66.43% | 55.44% | 59.98% | |
Mean test time (s) | 0.0032 | 00.29 | 00.30 | |
