Table 2 Comparison of the proposed model versus individual pyramidal models, highlighting the advantage of the proposed model in balancing sensitivity and specificity for superior accuracy. All values are presented as the average (Mean) with a measure of variability (Standard Deviation, SD). Significant values are in bold.
From: Precise grading of non-muscle invasive bladder cancer with multi-scale pyramidal CNN
ShuffleNet models | Sensitivity % | Specificity % | Accuracy % | F1-Score % |
|---|---|---|---|---|
Fusion (proposed) | 94.47 ± 0.93 | 94.03 ± 0.95 | 94.25 ± 0.70 | 94.29 ± 0.70 |
512-patch model | 94.52 ± 0.99 | 92.83 ± 1.14 | 93.68 ± 0.64 | 93.75 ± 0.62 |
256-patch model | 90.72 ± 0.89 | 78.90 ± 1.10 | 84.83 ± 0.69 | 85.71 ± 0.64 |
128-patch model | 56.16 ± 1.30 | 90.84 ± 1.10 | 73.44 ± 0.91 | 67.96 ± 1.19 |
