FractalAnalysis

class PVBM.FractalAnalysis.MultifractalVBMs(n_dim=10, n_rotations=25, optimize=True, min_proba=0.01, maxproba=0.98)

A class that can perform multifractal biomarker computation for a fundus vasculature segmentation.

Parameters:

• n_dim (int) – Maximum dimension used to calculate Singularity Length.

• n_rotations (int) – Number of rotations performed on the given segmentation from which the optimal run will be chosen.

• optimize (bool) – A boolean value that specify if the computation will be made using several rotated version of the segmentation.

• min_proba (float) – A minimum probability for the occupancy of calculated boxes (boxes with smaller probability will be ignored).

• max_proba (float) – A maximum probability for the occupancy of calculated boxes (boxes with higher probability will be ignored).

compute_multifractals(segmentation)

Computes the multifractal biomarkers of a given retinal vasculature segmentation, specifically the dimensions D0, D1, D2, and the singularity length.

Parameters:

segmentation (np.ndarray) – An (N,N) binary numpy array representing the segmentation to be analyzed.

Returns:

A numpy array containing the computed biomarkers [D0, D1, D2, Singularity Length].

Return type:

np.ndarray