MATLAB: 2-D Histogram Correlation Issue

2-d histogramcorrelationMATLABsimilaritysum of squared differences

Dear all,

I have about 500 distributions and trying to find which has highest correlation, which is most similar to my reference 2-D histogram. My reference distribution;

To select the most similar distribution, I'm currently using the sum of square differences method and the result is as follows;

But the result is not satisfying at the moment. I've checked with plotting many distributions and I think the following distribution is more similar to my reference distribution. The titles are the values of sum of squared differences, smaller one should indicate the higher correlation.

Could you please advice if I'm using the correct method to find the distribution most similar to reference one? What is your recommendation?

This is how I calculate sum of squared differences;

    sum(sum((map-referencemap).^2));

Thank you in advance!

Best Answer

It depends on how much you want to penalize the bin differences. The MSE will heavily penalize outliers much more than close values while the Median Absolute Difference or Mean Absolute Difference (see Wikipedia) has more of a linear weighting.

Related Solutions

MATLAB: I computed LBP histogram for two images in matlab, what method used to compare the LBP histograms

Hi there. The question is how to compare two histograms. In other words you want to find their similarity/dissimilarity measure. One way to do so is to calculate their SSD- Sum oF Square differences, which is a common way to present difference between tow feature vectors. To do so htry: diffHistLBP=hist1-hist2; SSD=sum(diffHistLBP.^2);

Naturally there are many other metrics... You can normalize/equalize the histogram before finding their SSD. Hope this is helpful, Nikolay

MATLAB: Too high correlation value from xcorr and corrcoef for uncorelated sequences

Ah, but there is quite a lot of correlation; remember it's not the magnitude of the values that matters, it's only whether they tend to "move together". To see this, try

figure, subplot(2,1,1)
yyaxis left, plot(dat(:,1))
yyaxis right,plot(dat(:,2))
ylim([0.35 1.5])

to overlay the two data series on top of each other at a scale factor that makes them roughly match each other in mean amplitude. What this shows is pretty much the same overall trend and that even some of the substructure is similar; particularly there being a drop towards the RH end around 5000. But, mostly what the correlation is measuring is the overall trend.

To see this,

subplot(2,1,2)
dtr=detrend(dat);
plot(dtr)
corr=corrcoef(dtr(:,1),dtr(:,2))
corr =
  1.0000    0.1071
  0.1071    1.0000

which is probably much more like what you were expecting.

Best Answer

Related Solutions

MATLAB: I computed LBP histogram for two images in matlab, what method used to compare the LBP histograms

MATLAB: Too high correlation value from xcorr and corrcoef for uncorelated sequences

Related Question