Table 4.
Comparative WBC detection among methods which considers the complete data set of 30 images corrupted by different levels of Gaussian noise.
Noise level | Method | Leukocytes detected | Missing | False alarms | DR | FAR |
---|---|---|---|---|---|---|
σ = 5 Gaussian noise 517 leukocytes | BSV [3] | 214 | 303 | 98 | 41.39% | 18.95% |
IO [4] | 366 | 151 | 87 | 70.79% | 16.83% | |
Wang [5] | 358 | 159 | 84 | 69.25% | 16.25% | |
GAB [12] | 407 | 110 | 76 | 78.72% | 14.70% | |
DE-based | 487 | 30 | 21 | 94.20% | 4.06% | |
| ||||||
σ = 10 Gaussian noise 517 leukocytes | BSV [3] | 162 | 355 | 129 | 31.33% | 24.95% |
IO [4] | 331 | 186 | 112 | 64.02% | 21.66% | |
Wang [5] | 315 | 202 | 124 | 60.93% | 23.98% | |
GAB [12] | 363 | 154 | 113 | 70.21% | 21.86% | |
DE-based | 471 | 46 | 35 | 91.10% | 6.77% |