Research Article
Robust Method for Semantic Segmentation of Whole-Slide Blood Cell Microscopic Images
Table 4
Quantitative comparison of proposed technique with existing techniques.
| Paper | WBCs seg | RBCs seg | Platelets seg | Accuracy (%) | Global accuracy | Evaluation matrices | Cell type |
| Nikitaev et al. [4] | ✓ | ✕ | ✕ | WBCs = 82 | ✕ | Accuracy | Single cell | Miao and Xiao [5] | ✓ | ✓ | ✕ | WBCs = 97.2 RBCs = 94.8 | ✕ | Accuracy | Single cell | Cao et al. [19] | ✓ | ✕ | ✕ | RBCs = 93.75 | ✕ | Accuracy | Single cell | Shahin et al [15] | ✓ | ✕ | ✕ | WBCs = 97.6 | ✕ | Accuracy | Single cell | Mohammed et al. [29] | ✓ | ✕ | ✕ | WBCs = 90 | ✕ | Accuracy | Single cell | Proposed method | ✓ | ✓ | ✓ | RBCs = 97.45 WBCs = 93.34 Platelets = 85.11 | 97.18% | Accuracy, IoU, mean and weighted IoU, BFScore | Whole-slide image, segment WBCs, RBCs, and platelets simultaneously |
|
|