Coronary Artery Segmentation on Coronary artery segmentation dataset (5-fold cross-val)
97.55Accuracy (Acc)DenseNet201_UNet++
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| DenseNet201_UNet++Pre-processing=Proposed (3-Ch), Trainable Parameters (Million)=48.6, Average Inference time (s)=1.1442025.10 | 97.55 | 61.12 | 75.87 | 75.31 | 76.52 | 98.67 | 23.48 | 1.33 | |
| DenseNet121_UNetPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=13.6, Average Inference time (s)=0.6812025.10 | 97.54 | 60.67 | 75.51 | 75.74 | 75.33 | 98.72 | 24.66 | 1.28 | |
| Resnet50_MAnetPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=147.4, Average Inference time (s)=0.4882025.10 | 97.53 | 60.47 | 75.37 | 75.63 | 75.17 | 98.71 | 24.83 | 1.29 | |
| Proposed networkPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=26.9, Average Inference time (s)=0.6962025.10 | 97.52 | 61.43 | 76.1 | 73.79 | 78.61 | 98.52 | 21.39 | 1.48 | |
| Resnet50_Self-ONN_Unet q3Pre-processing=Proposed (3-Ch), Trainable Parameters (Million)=50.53, Average Inference time (s)=0.4022025.10 | 97.52 | 60.51 | 75.4 | 75.52 | 75.44 | 98.69 | 24.56 | 1.31 | |
| Inception_v4_Unet++Pre-processing=Proposed (3-Ch), Trainable Parameters (Million)=59.3, Average Inference time (s)=0.7692025.10 | 97.51 | 60.84 | 75.65 | 74.56 | 76.84 | 98.6 | 23.16 | 1.4 | |
| Densenet121_LinknetPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=10.4, Average Inference time (s)=0.7022025.10 | 97.49 | 60.51 | 75.39 | 74.32 | 76.52 | 98.6 | 23.48 | 1.4 | |
| Densenet121_Unet++_scSEPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=14.2, Average Inference time (s)=1.3092025.10 | 97.48 | 61.18 | 75.92 | 74.52 | 77.36 | 98.58 | 22.58 | 1.38 | |
| Efficientnet_b2_Unet++Pre-processing=Proposed (3-Ch), Trainable Parameters (Million)=10.4, Average Inference time (s)=0.6402025.10 | 97.48 | 60.35 | 75.27 | 74.38 | 76.24 | 98.6 | 23.76 | 1.4 | |
| MedSAMPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=93.7, Average Inference time (s)=1.5122025.10 | 97.43 | 58.2 | 73.56 | 76.12 | 71.4 | 98.8 | 28.6 | 1.2 | |
| Proposed networkPre-processing=Ben Graham, Trainable Parameters (Million)=26.9, Average Inference time (s)=0.6942025.10 | 97.41 | 60.77 | 75.59 | 71.87 | 79.77 | 98.34 | 20.23 | 1.66 | |
| Proposed networkPre-processing=N/A, Trainable Parameters (Million)=26.9, Average Inference time (s)=0.6982025.10 | 97.4 | 60.18 | 75.14 | 72.56 | 78.11 | 98.42 | 21.89 | 1.58 | |
| Proposed networkPre-processing=CLAHE, Trainable Parameters (Million)=26.9, Average Inference time (s)=0.6922025.10 | 97.35 | 60.04 | 75.03 | 71.77 | 78.86 | 98.33 | 21.14 | 1.67 | |
| mit_b1_FPNPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=15.03, Average Inference time (s)=0.6322025.10 | 97.27 | 58.73 | 74 | 71.06 | 77.24 | 98.33 | 22.76 | 1.67 | |
| MMDC-NetPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=19.36, Average Inference time (s)=1.5542025.10 | 97 | 60.95 | 75.39 | 65.32 | 89.91 | 97.37 | 10.09 | 2.63 | |
| TiM-NetPre-processing=Proposed (3-Ch), Trainable Parameters (Million)=104.06, Average Inference time (s)=1.8342025.10 | 96.97 | 43.73 | 56.52 | 84.5 | 48.71 | 99.52 | 51.29 | 0.48 |