Table 7.
Features | Magnification | Image_accuracy (%) | Patient_accuracy (%) | Sensitivity (%) | Precision (%) | F1-score (%) |
---|---|---|---|---|---|---|
GLCM1+APVEC | 40X | 94.22 ± 2.18 | 93.70 ± 2.68 | 95.77 ± 2.19 | 96.39 ± 1.57 | 96.07 ± 1.77 |
100X | 92.78 ± 2.88 | 91.80 ± 3.67 | 95.47 ± 4.45 | 94.83 ± 1.84 | 95.09 ± 2.29 | |
200X | 94.77 ± 2.18 | 94.33 ± 2.99 | 97.23 ± 3.64 | 95.89 ± 2.13 | 96.50 ± 1.65 | |
400X | 91.20 ± 2.17 | 91.59 ± 1.94 | 91.82 ± 3.90 | 96.03 ± 2.9 | 93.80 ± 1.80 | |
GLCM1+ HIM | 40X | 94.13 ± 1.85 | 93.15 ± 2.52 | 95.26 ± 2.24 | 96.75 ± 1.33 | 95.99 ± 1.56 |
100X | 92.69 ± 2.71 | 91.17 ± 3.58 | 95.46 ± 4.74 | 94.75 ± 1.71 | 95.03 ± 2.22 | |
200X | 94.88 ± 2.39 | 94.30 ± 3.17 | 96.98 ± 4.00 | 96.29 ± 2.47 | 96.56 ± 1.76 | |
400X | 91.19 ± 2.73 | 91.54 ± 2.26 | 91.55 ± 4.75 | 96.26 ± 2.44 | 93.76 ± 2.27 | |
GLCM1+Wavelet | 40X | 93.24 ± 3.45 | 92.85 ± 3.75 | 94.67 ± 3.7 | 96.06 ± 2.14 | 95.35 ± 2.61 |
100X | 92.37 ± 3.47 | 91.38 ± 4.29 | 95.03 ± 4.67 | 94.66 ± 1.98 | 94.80 ± 2.73 | |
200X | 94.25 ± 1.64 | 93.47 ± 2.39 | 96.35 ± 3.45 | 96.06 ± 2.33 | 96.15 ± 1.27 | |
400X | 90.62 ± 1.45 | 90.35 ± 1.52 | 91.49 ± 3.33 | 95.59 ± 3.08 | 93.42 ± 1.27 | |
GLCM1+Tamura | 40X | 93.76 ± 2.89 | 93.35 ± 3.24 | 94.81 ± 3.14 | 96.64 ± 1.80 | 95.71 ± 2.35 |
100X | 92.28 ± 3.73 | 91.07 ± 4.44 | 95.63 ± 5.63 | 94.04 ± 1.97 | 94.75 ± 2.95 | |
200X | 94.88 ± 1.89 | 94.45 ± 2.90 | 97.36 ± 2.98 | 95.98 ± 2.71 | 96.61 ± 1.30 | |
400X | 90.97 ± 1.76 | 91.26 ± 1.84 | 92.23 ± 3.63 | 95.34 ± 2.91 | 93.69 ± 1.52 | |
GLCM1+LBP | 40X | 88.02 ± 3.43 | 88.28 ± 3.79 | 90.60 ± 3.33 | 93.33 ± 3.70 | 91.83 ± 2.82 |
100X | 89.03 ± 4.85 | 89.12 ± 4.08 | 93.05 ± 2.69 | 92.30 ± 5.04 | 92.64 ± 3.45 | |
200X | 89.71 ± 3.10 | 89.26 ± 3.22 | 92.49 ± 3.25 | 93.68 ± 2.35 | 93.06 ± 2.33 | |
400X | 87.18 ± 4.14 | 86.42 ± 3.79 | 89.18 ± 3.76 | 93.08 ± 5.15 | 90.99 ± 3.15 | |
GLCM1+CLBP | 40X | 93.86 ± 2.62 | 93.07 ± 3.55 | 95.69 ± 4.12 | 96.03 ± 2.36 | 95.84 ± 2.14 |
100X | 91.72 ± 3.07 | 90.46 ± 3.82 | 94.06 ± 3.75 | 94.70 ± 1.62 | 94.36 ± 2.39 | |
200X | 94.34 ± 1.99 | 93.51 ± 2.61 | 97.48 ± 1.60 | 95.15 ± 2.21 | 96.29 ± 1.29 | |
400X | 89.66 ± 3.23 | 88.94 ± 3.64 | 90.87 ± 5.22 | 94.83 ± 3.75 | 92.69 ± 2.66 | |
GLCM1+Gabor | 40X | 87.63 ± 4.43 | 87.22 ± 4.8 | 93.86 ± 2.38 | 90.47 ± 6.57 | 91.92 ± 3.15 |
100X | 90.28 ± 2.04 | 89.95 ± 2.72 | 93.44 ± 2.97 | 93.71 ± 3.22 | 93.51 ± 1.34 | |
200X | 91.28 ± 2.64 | 89.81 ± 4.09 | 97.06 ± 1.88 | 91.92 ± 3.66 | 94.37 ± 1.70 | |
400X | 89.02 ± 2.83 | 87.80 ± 2.54 | 96.72 ± 1.77 | 89.22 ± 3.37 | 92.79 ± 1.96 | |
GLCM1+Hog | 40X | 85.55 ± 5.78 | 86.56 ± 5.32 | 93.02 ± 2.53 | 88.39 ± 6.82 | 90.56 ± 4.20 |
100X | 86.62 ± 5.08 | 88.34 ± 3.18 | 94.06 ± 3.22 | 89.23 ± 7.32 | 91.35 ± 3.14 | |
200X | 90.47 ± 3.78 | 90.04 ± 3.61 | 93.68 ± 4.40 | 93.55 ± 2.85 | 93.57 ± 2.92 | |
400X | 88.15 ± 4.05 | 88.13 ± 3.72 | 91.45 ± 2.54 | 92.28 ± 4.56 | 91.82 ± 3.01 |
The bold values indicate that the recognition accuracy of combining the two features is better than that of a single feature.