scalar = normalizedCrossCorrelationMsr3d (inImg3d1,inImg3d2)

Detailed Description

measures the global Normalized Cross-Correlation between two 3D images

This algorithms computes the global Normalized Cross-Correlation (NCC) between the two input images. In opposition to the Normalized Cross-Correlation 3d algorithm which returns an image, the result of this algorithm is a a single scalar value.

The 3D NCC is calculated according to the following formula :

$NCC(InImg3d1(\textbf{x}), InImg3d2(\textbf{x})) = \frac{\sum_{\textbf{x} \in \Omega} (InImg3d1(\textbf{x}) - \overline{InImg3d1}) (InImg3d2(\textbf{x}) - \overline{InImg3d2}) } {\sqrt{\sum_{\textbf{x} \in \Omega} (InImg3d1(\textbf{x}) - \overline{InImg3d1})^2 \sum_{\textbf{x} \in \Omega} (InImg3d2(\textbf{x}) - \overline{InImg3d2})^2}}$

Where $\Omega = (sizeX, sizeY, sizeZ)$ is the image domain definition.

The result belongs to $[-1, 1]$ , and equals 1 when the images exactly matches, -1 when the images are exactly inverted and 0 when the images are totally different from each other.

If the input images has several slices along the C or T dimensions, the result is not a scalar but a collection of ipsdk::imaproc::attr::NCCResult. This data structure stores the actual result in its Result field. The other fields can be ignored as long as they are used for the calculation.

Please, see Normalized Cross-Correlation Measure 2d for an illustration of the global NCC in 2D case.

Example of Python code :

Example imports

import PyIPSDK

import PyIPSDK.IPSDKIPLGlobalMeasure as glbmsr

Code Example

    # retrieve measurement results
    nccValue = glbmsr.normalizedCrossCorrelationMsr3d(inImg1, inImg2)
    print("Global NCC value =  " + str(nccValue))

Example of C++ code :

Example informations

Header file

#include <IPSDKIPL/IPSDKIPLGlobalMeasure/Processor/NormalizedCrossCorrelationMsr3d/NormalizedCrossCorrelationMsr3d.h>

Code Example

const ipReal64 ncc = normalizedCrossCorrelationMsr3d(pInImg1, pInImg2);