In the case of Random Forest with no hyperparameter tuning, the accuracy increased by 4.6% [95% CI: 4.1 to 5.2%] when using WLIs as opposed to 5.6% [95% CI: 5.1 to 6.1%] when using SAVE images, though this difference did not reach statistical significance in the aggregate count comparison [ p = 0.120].
← all excerpts
Efficacy of Spectral-Aided Visual Enhancer in Classification of Esophageal Cancer.
1
0.1200
0.1200