How to quantify the object shift between projection images using Matlab?

Hi, I am tasked with quantifying the shift between these 2D simulated images, by implementing an algorithm (e.g. cross-correlation or centre of mass) using matlab. I have attached the data and script for the images. Any help would be appreciated.

[Yanique]

@Image Analyst please help :)

Answer 1

[Image Analyst]

I don't see any images attached, but try imregister or imregcorr

Or try normalized cross correlation. See attached demo.

[Yanique]

I included the Matlab script with the image data. And the excel data files.

I will try imregister and imregcorr. Is it likely that they return a constant instead of an image to quantity a shift in the y direction?

[Yanique]

Hi again @Image Analyst , I'm new to matlab and cannot figure out how to use the suggested imregister or imregcorr for this application. I have edited the question to include images of the shift a dot was placed at the centre of the graph (75 , 75). I need a value in pixels or cm to verify that the image moved. Please help

[Image Analyst]

Now that I can see your images, why don't you just threshold and find the centroid of the blob? Do this for each image and see how the centroid shifted. Something like

mask1 = grayImage1 > someThresholdValue;
% Take largest blob only.
mask1 = bwareafilt(mask1, 1);
props1 = regionprop(mask1, 'Centroid');
xy1 = props1.Centroid
% Now for image 2
mask2 = grayImage2 > someThresholdValue;
% Take largest blob only.
mask2 = bwareafilt(mask2, 1);
props2 = regionprop(mask2, 'Centroid');
xy2 = props2.Centroid
% Now compute delta x and y
deltax = xy1(1) - xy2(1)
deltay = xy1(2) - xy2(2)

See my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image to find blobs, and then measure things about the blobs, and extract certain blobs based on their areas or diameters.

[Yanique]

Thank you so much for your help, this looks like what I need. I have a problem converting my data to grayscale. The image plot I have is a matlab.primitive.image with cannot be converted. Is there another way to plot this 2D image with x and y: 0-290 and z containing intensity values. I found this mesgrid code but I cannot do any analysis on the image afterward. @Image Analyst

Thank you for all your help

C = readmatrix('[e-mail]');
D = readmatrix('[e-mail]');

xC = C(:,1);
yC = C(:,2);
zC = C(:,3);

xD = D(:,1);
yD = D(:,2);
zD = D(:,3);

n = 290.39;

[Xc, Yc] = meshgrid(linspace(min(-150),max(150),n), linspace(min(-150),max(150),n));
Zc = griddata(xC,yC,zC,Xc,Yc);

[Xd, Yd] = meshgrid(linspace(min(-150),max(150),n), linspace(min(-150),max(150),n));
Zd = griddata(xD,yD,zD,Xd,Yd);

figure (1) % Plot the original image C
GlobalPosZX_C = imagesc(Zc);
title('O R I G I N A L  I M A G E: Unshifted')
centre_1 = drawline('Position',[145.175 145.175;145.175 145.175],'StripeColor','r');

figure (2) % Plot the shifted image 5cm
GlobalPosZX_D = imagesc(Zd);
title('S H I F T E D : 5cm Shift')
centre_4 = drawline('Position',[145.175 145.175;145.175 145.175],'StripeColor','r');