Table 1 Geometry features extracted from a MC boundary
Feature index | Feature | Note |
|---|---|---|
GF1 | Area of the MC | Number of pixels in the segmented region |
GF2 | Perimeter of the MC | Number of pixels on the border of the MC |
GF3 | Compactness | Calculated with \( C=1-\frac{4\pi \times \left(\mathrm{area}\right)}{{\left(\mathrm{perimeter}\right)}^2} \) |
GF4 | NDM2 | Normalized distance-based moments |
GF5 | NDM3 | |
GF6 | NDM4 | |
GF7 | Fourier feature | Fourier feature calculated taking boundary pixel as a complex number |
GF8 | NRL mean | Statistical values from normalized radial length |
GF9 | NRL standard deviation | |
GF10 | NRL entropy | |
GF11 | NRL area ratio | |
GF12 | RGO mean | Statistical values from relative gradient orientation, measure spiculation, and defined as the acute angle θ between radial direction of a point on the contour and the gradient direction of the point |
GF13 | RGO standard deviation | |
GF14 | RGO entropy |