volume_stats.py

"""
Contains various functions for computing statistics over 3D volumes
"""
import numpy as np

def Dice3d(a, b):
    """
    This will compute the Dice Similarity coefficient for two 3-dimensional volumes
    Volumes are expected to be of the same size. We are expecting binary masks -
    0's are treated as background and anything else is counted as data

    Arguments:
        a {Numpy array} -- 3D array with first volume
        b {Numpy array} -- 3D array with second volume

    Returns:
        float
    """
    if len(a.shape) != 3 or len(b.shape) != 3:
        raise Exception(f"Expecting 3 dimensional inputs, got {a.shape} and {b.shape}")

    if a.shape != b.shape:
        raise Exception(f"Expecting inputs of the same shape, got {a.shape} and {b.shape}")

    a1 = a>0
    b1 = b>0
    
    intersection = np.sum(a1*b1)
    volumes = np.sum(a1) + np.sum(b1)
    
    if volumes == 0:
        return -1
    
    return 2.*float(intersection)/ float(volumes)

def Jaccard3d(a, b):
    """
    This will compute the Jaccard Similarity coefficient for two 3-dimensional volumes
    Volumes are expected to be of the same size. We are expecting binary masks - 
    0's are treated as background and anything else is counted as data

    Arguments:
        a {Numpy array} -- 3D array with first volume
        b {Numpy array} -- 3D array with second volume

    Returns:
        float
    """
    if len(a.shape) != 3 or len(b.shape) != 3:
        raise Exception(f"Expecting 3 dimensional inputs, got {a.shape} and {b.shape}")

    if a.shape != b.shape:
        raise Exception(f"Expecting inputs of the same shape, got {a.shape} and {b.shape}")

    a1 = a>0
    b1 = b>0
    
    intersection = abs(np.sum(a1*b1))
    volumes = abs(np.sum(a1)) + abs(np.sum(b1)) - intersection
    
    return float(intersection)/float(volumes)