Table 3. MSB classification performance of fused features.
Features | Magnification | Image_accuracy(%) | Patient_accuracy(%) | Sensitivity(%) | Precision(%) | F1_score(%) |
---|---|---|---|---|---|---|
Average pool_1+GLCM | 40× | 93.26±2.73 | 93.30±3.12 | 94.13±2.95 | 90.98±5.60 | 96.71±2.28 |
100× | 92.78±1.90 | 93.03±3.18 | 93.64±2.86 | 90.88±7.78 | 96.62±2.95 | |
200× | 95.06±1.57 | 94.38±1.49 | 95.78±2.23 | 92.86±5.52 | 97.59±1.80 | |
400× | 92.08±2.74 | 92.11±2.97 | 95.47±2.89 | 88.63±9.00 | 95.67±3.66 | |
Average pool_2+GLCM | 40× | 94.86±3.02 | 94.11±3.38 | 96.93±3.17 | 93.07±5.37 | 97.45±2.15 |
100× | 94.61±2.22 | 94.12±2.50 | 94.61±2.70 | 95.81±2.33 | 98.42±1.08 | |
200× | 94.99±2.00 | 95.09±2.35 | 96.53±2.01 | 94.42±5.48 | 98.08±1.83 | |
400× | 92.65±2.32 | 92.42±2.12 | 95.33±1.91 | 87.87±6.64 | 95.41±2.83 | |
Average pool_3+GLCM | 40× | 95.02±1.97 | 94.92±2.20 | 97.01±2.66 | 94.16±3.96 | 98.01±1.32 |
100× | 94.91±1.99 | 94.60±2.32 | 97.14±1.83 | 91.70±3.52 | 97.09±1.57 | |
200× | 95.62±1.98 | 95.21±2.08 | 97.19±2.20 | 91.96±3.18 | 97.32±1.15 | |
400× | 92.35±2.26 | 91.90±3.08 | 93.03±2.97 | 90.44±3.21 | 96.31±1.51 | |
Global average pool+GLCM | 40× | 95.02±1.87 | 94.53±2.50 | 97.22±1.20 | 92.47±4.43 | 97.44±1.49 |
100× | 94.75±2.62 | 94.05±3.21 | 96.25±2.61 | 92.69±3.37 | 97.39±1.50 | |
200× | 95.38±1.89 | 95.01±2.23 | 96.68±2.91 | 96.90±2.54 | 98.96±0.86 | |
400× | 92.91±2.41 | 92.20±2.62 | 93.64±2.81 | 93.99±4.09 | 97.56±1.91 | |
FC+GLCM | 40× | 94.36±2.02 | 94.17±2.38 | 96.70±2.26 | 91.23±4.48 | 97.05±1.46 |
100× | 94.13±1.99 | 93.45±2.68 | 96.83±2.68 | 85.56±5.59 | 95.35±1.55 | |
200× | 95.55±1.22 | 95.35±1.41 | 97.16±1.67 | 91.84±4.95 | 97.31±1.65 | |
400× | 92.13±2.54 | 91.42±2.43 | 93.65±2.08 | 88.07±3.84 | 95.40±2.12 | |
block4+GLCM | 40× | 94.32±2.76 | 93.80±2.78 | 96.11±2.32 | 96.19±1.95 | 96.14±2.03 |
100× | 94.31±2.72 | 94.27±2.90 | 96.36±2.59 | 95.91±1.77 | 96.13±2.08 | |
200× | 95.52±1.77 | 95.28±1.81 | 96.74±1.74 | 97.29±2.46 | 96.99±1.28 | |
400× | 92.77±2.50 | 92.53±2.40 | 95.00±1.83 | 95.05±2.75 | 95.01±1.91 | |
block6+GLCM | 40× | 94.09±1.81 | 93.66±1.88 | 95.93±1.30 | 96.12±1.88 | 96.02±1.35 |
100× | 93.60±2.24 | 93.71±2.40 | 95.75±2.22 | 95.59±1.66 | 95.67±1.79 | |
200× | 95.36±1.92 | 95.07±1.76 | 97.40±1.43 | 96.44±2.25 | 96.90±1.39 | |
400× | 92.90±1.98 | 92.78±2.08 | 94.23±2.33 | 95.95±2.17 | 95.06±1.61 | |
block14+GLCM | 40× | 96.75±1.96 | 96.33±2.14 | 97.86±1.53 | 97.76±2.03 | 97.80±1.41 |
100× | 95.20±2.31 | 95.26±2.60 | 95.79±2.11 | 97.63±1.65 | 96.69±1.78 | |
200× | 96.29±1.49 | 96.09±1.79 | 97.14±1.77 | 97.86±1.29 | 97.49±1.12 | |
400× | 93.15±2.30 | 92.99±2.85 | 94.43±2.51 | 96.09±2.42 | 95.23±1.83 | |
block19+GLCM | 40× | 94.86±2.31 | 94.50±2.64 | 95.93±2.39 | 97.10±1.38 | 96.51±1.69 |
100× | 94.87±2.11 | 95.07±2.19 | 96.53±1.49 | 96.54±2.37 | 96.53±1.61 | |
200× | 95.79±1.71 | 95.48±1.47 | 97.81±1.48 | 96.64±2.48 | 97.20±1.19 | |
400× | 93.02±2.35 | 92.99±2.39 | 95.07±2.77 | 95.34±2.51 | 95.17±1.84 | |
block22+GLCM | 40× | 96.29±1.82 | 95.87±2.11 | 98.52±1.68 | 96.53±1.43 | 97.51±1.32 |
100× | 94.93±2.27 | 94.90±2.35 | 96.57±1.49 | 96.56±2.04 | 96.56±1.71 | |
200× | 95.73±2.09 | 95.48±2.16 | 96.77±2.21 | 97.48±1.80 | 97.11±1.54 | |
400× | 93.09±2.31 | 92.86±2.23 | 94.19±2.60 | 96.22±2.06 | 95.17±1.83 | |
block23+GLCM | 40× | 95.86±1.73 | 95.49±1.51 | 96.90±1.90 | 97.48±1.38 | 97.18±1.31 |
100× | 95.21±2.18 | 95.23±2.61 | 95.99±2.30 | 97.47±1.68 | 96.71±1.69 | |
200× | 96.57±1.82 | 96.05±1.94 | 97.40±2.35 | 97.97±1.47 | 97.67±1.35 | |
400× | 93.05±2.05 | 92.63±2.82 | 93.98±2.50 | 96.40±2.41 | 95.14±1.67 |