Table 7 State of Art comparison with the proposed FasNet model with attention and Monte Carlo dropout.
Year/ Ref | Technique | Dataset Used | Dice Coefficient | Jaccard Index (IOU) |
|---|---|---|---|---|
2024/ Chen et al.21 | Multi-Scale Liver Tumor Segmentation Algorithm | LiTS17 | 74.8% | ------- |
2024/ You et al.22 | Parallel Graph Convolutional Network (PGC-Net) | LiTS17 3DIRCADb | 73.63% 74.16% | ------- |
2024/ He et al.23 | PAKS-Net | KiTS19 LiTS17 Pancreas LOTUS | 89.3% 76.9% 59.8% 73.8% | ------- |
2024/ Shui et al.24 | MSFF, MFF, EI, EG modules. | LiTS17 3Dircadb | 85.55% 80.14% | 81.11% 81.68% |
2024/ Biswas et al.25 | GAN-driven data augmentation strategy | LiTS17 3Dircadb MIDAS | 60.5% 87.2% 90.8% | ------- |
2023/ Wang et al.26 | Context Fusion Network includes Twin-Split Attention (TSA) modules and Multi-Scale-Aware (MSA) skip connections. | LiTS17 3Dircadb | 85.97% 80.11% | 81.56% 83.67% |
Proposed Model | Hybrid “FasNet” Model with Attention and Monte Carlo Dropout | LiTS17 | 87.66% | 84.87% |
