Compare two histograms with Correlation Coefficient

Im am trying to get a histogram of the hue of an image and then compare two different image histograms with correlation coefficient. "Corr2" is not working..

img = imread('hb1.jpeg'); hsvImage = rgb2hsv(img); hImage = hsvImage(:, :, 1); figure; subplot(2,2,1); hHist = histogram(hImage);

img = imread('hb2.jpeg'); hsvImage = rgb2hsv(img); hImage = hsvImage(:, :, 1); figure; subplot(2,2,1); hHist2 = histogram(hImage);

R=corr2(hHist, hHist2);

Answer 1

[Steven Lord]

The corrcoef function cannot accept histogram object handles. You may have better luck using corrcoef on the vectors of bin counts from two calls to histcounts (using the same set of edges for each call) instead.

% Sample data
x1 = randn(1, 1e5);
x2 = 2*randn(1, 1e5);

% Obtain the vectors of bin counts. Use the set of edges returned from the first call
% as an input to the second call so the vector of bin counts are the same length and
% measure "apples to apples"
[n2, edges] = histcounts(x2);
n1 = histcounts(x1, edges);

% Compute the correlation coefficient between the vectors of bin counts
corrcoef(n1, n2)

You might also find this Stack Exchange discussion interesting and/or informative.

Answer 2

[Changoleon]

Why it does not work? What is the error?

How about you try this:

R=corrcoef(hHist, hHist2);

[Sandra Marta]

This is the error: Expected input number 1, A, to be one of these types:

logical, numeric

Instead its type was matlab.graphics.chart.primitive.Histogram.