Label Quantification¶

This algorithm takes in a labelled image and calculates several characteristics for each label. Below are some of the characteristics

Volume by voxel counts
Equivalent sphere diameter by voxel counts
Bounding box diagonal
Principal Component Analysis (PCA)
Ellipsoid fitting by PCA
Equivalent circle diameter by PCA
Isosurface by marching cude.
Surface area
Surface Volume
Equivalent sphere diameter from surface volume
Sphercity
Normalised surface area to volume ratio (Radius*Sa/Vol)

The details of this algorithm are given in Evaluation of 3D bioactive glass scaffolds dissolution in a perfusion flow system with X-ray microtomography. Yue S, Lee PD, Poologasundarampillai G, Jones JR. DOI: 10.1016/j.actbio.2011.02.009

API¶

LabelQuantification(input_3d_volume, origin, voxel_size, min_data_value, max_data_value, minimum_feature_size)

input_3d_volume: numpy array of 3d volume. data type reflects the function suffix.
origin: numpy array with 3 values.
voxel_size: numpy array representing a voxel size in three dimensions (x,y,z)
min_data_value: minimum value of the input volume
max_data_value: maximum value of the input volume
minimum_feature_size: minimum size of the feature

returns: an list with feature names and another numpy array with each row representing a label and it corresponding values.

Example¶

To run the example code you need to download the following data file: Foam Data

To run the below code you need to install the following packages:

tifffile ( conda install -c conda-forge tifffile )

# Imports
from ccpi.quantification.LabelQuantification import LabelQuantification
import numpy as np
import math
from tifffile import TiffFile

    #Read the input volume which is labelled during segmentation
img = TiffFile('FoamData.tif')
data = img.asarray()

    #voxel size
voxel_size = np.ones(3,dtype=np.float32)
    #origin
origin = np.zeros(3,dtype=np.float32)

    #computes the 3d quantification
lqsNames, lqsValues = LabelQuantification(data, origin, voxel_size, float(np.amin(data)), float(np.amax(data)), 100.0)
lqs.compute()

print(lqsNames)
img.close()