Table 2 Results of the comparison of the proposed method with other segmentation methods from the literature
From: Modified distance regularized level set evolution for brain ventricles segmentation
Methods | Sensitivity | Specificity | Accuracy |
|---|---|---|---|
Proposed method (atrophy CT, n = 20; normal CT, n = 20; atrophy MRI, n = 20; normal MRI, n = 20) | 65%–90% | 98%–99% | 95%–98% |
DRLSE (implemented) | 74%–79% | 91%–93% | 92%–94% |
Region based level set (MRI, n = 124) [38] | 84%–89% | 96%–97% | – |
DRLSE and deep structure (MRI, n = 45) [39] | – | – | 84%–95% |
Fuzzy and evidential reasoning (MRI, n = 17) [40] | 55%–90% | – | 60%–80% |
Hybrid genetic and EM algorithm (MRI, n = 10) [41] | – | – | 45%–85% |
Bayesian morphometry (MRI, n = 95) [42] | 60%–80% | 30%–60% | – |
EGS-SOM (MRI, n = 57) [43] | 76%–80% | 95%–96% | – |
HFS-SOM (MRI, n = 57) [43] | 66%–77% | 80%–85% | – |