Table 3 Performance evaluations of Ranklet , GLCM , LBP and PCBP with bootstrap method (mean ± standard deviation)
From: Robust phase-based texture descriptor for classification of breast ultrasound images
Methods | AUC | ACC (%) | SENS (%) | SPEC (%) | PPV (%) | NPV (%) | MCC (%) |
|---|---|---|---|---|---|---|---|
Ranklet | 0.843 ± 0.044 | 80.68 ± 5.55 | 76.68 ± 10.86 | 85.39 ± 7.24 | 83.45 ± 7.90 | 78.61 ± 9.57 | 0.625 ± 0.105 |
GLCM | 0.832 ± 0.043 | 78.40 ± 4.80 | 75.17 ± 11.35 | 82.36 ± 10.12 | 81.67 ± 8.83 | 77.39 ± 9.11 | 0.583 ± 0.094 |
LBP | 0.807 ± 0.048 | 77.77 ± 5.53 | 76.27 ± 10.03 | 79.87 ± 9.26 | 79.10 ± 8.95 | 77.61 ± 8.50 | 0.564 ± 0.108 |
PCBP | 0.862 ± 0.037 | 83.17 ± 4.81 | 83.36 ± 7.64 | 83.42 ± 8.32 | 84.25 ± 7.24 | 83.58 ± 7.23 | 0.670 ± 0.094 |