Table 12 AUC scores of TGV (or T2) under different combinations of gradient magnitude, gradient azimuth and gradient polar angle using leave-one-out cross validation
From: Vector textures derived from higher order derivative domains for classification of colorectal polyps
\({{\varvec{Q}}}_{{\varvec{g}}}^{{\varvec{m}}}\) | \({{\varvec{Q}}}_{{\varvec{g}}}^{{\varvec{a}}}\) | \({{\varvec{Q}}}_{{\varvec{g}}}^{{\varvec{p}}}\) | AUC score |
|---|---|---|---|
1 | 10 | 5 | 0.954 |
2 | 10 | 5 | 0.948 |
3 | 10 | 5 | 0.921 |
4 | 10 | 5 | 0.924 |
1 | 10 | 4 | 0.956 |
2 | 10 | 4 | 0.970 |
3 | 10 | 4 | 0.953 |