plot gaussian and standard deviation on my histogram

Hi everyone,

I have this code and have plotted the histogram and now want to have gaussian and standard deviation.

can anyone help me please?

figure('Name','A3{1}.BMd_Img_o_part1')

edges= linspace(0,255,21);

histogram(image1, 'BinEdges',edges,'normalization','pdf')

Answer 1

[Star Strider]

I would use normfit and normpdf, although there are also other options:

image1 = 42*randn(1,1000)+42*3;                                     % Create Data
figure('Name','A3{1}.BMd_Img_o_part1')
edges= linspace(0,255,21);
[muHat,sgHat] = normfit(image1);
pdfplot = normpdf(edges, muHat, sgHat);

figure
histogram(image1, 'BinEdges',edges,'normalization','pdf')
hold on
plot(edges, pdfplot, '-r', 'LineWidth',2)
hold off

.

[Fatemeh Shomal Zadeh]

Thanks so much for your help. Just one question, does this code has standard deviation as well?

[Fatemeh Shomal Zadeh]

I am sorry, my plot looks like this how I can increase the hight of the gaussian peak to reach to my histogram figure?

Also, how can I add the standard deviation to my figure?

Thanks a lot

[Star Strider]

As always, my pleasure!

Increasing the height would not be appropriate. Note that the plot overestimates some of the bins, and underestimates some of the others.

I am not certain what you intend by ’Also, how can I add the standard deviation to my figure?‘ One possibility would be to use a text object. Otherwise, it depends on what you want to do and how you want the standard deviation depicted. The standard deviation is the second normfit output, ‘sgHat’.

[Fatemeh Shomal Zadeh]

Thanks a lot for your explanatin.

My image1 data is the 368x315x3 uint8.

I do not understand this line image1 = 42*randn(1,1000)+42*3.

Can you help me how I can consider my data?

[Star Strider]

As always, my pleasure!

‘My image1 data is the 368x315x3 uint8.’

I do not have your data to work with, or the code you used to create ‘image1’.

‘I do not understand this line image1 = 42*randn(1,1000)+42*3.’

I created data to use to test my code. I always do this if the actual data are not provided (as they were not provided here).

Note that the full line is:

image1 = 42*randn(1,1000)+42*3;                                     % Create Data

.

[Fatemeh Shomal Zadeh]

This data was extracted from ultrasound Machine and as I have mentioned is 368x315x3 uint8 array.

Hope this information can help you and I would be gratfeul how I can plot the gaussian in this situation.

My code are listed as below;

imshow(Data_SWS_BMd_A3{1}.BMd_Img_o)

impixelinfo

xmin = 5;

xmax = 314;

ymin = 136;

ymax = 238;

imshow(Data_SWS_BMd_A3{1}.BMd_Img_o(ymin:ymax,xmin:xmax));

impixelinfo

image= Data_SWS_BMd_A3{1}.BMd_Img_o(ymin:ymax,xmin:xmax)

[n,m]=size(image)

id=fix(m/3)

image1=image(:,1:id);

image2=image(:,id+1:2*id);

image3=image(:,2*id+1:m);

figure

imshow(image1)

impixelinfo

figure

imshow(image2)

impixelinfo

figure

imshow(image3)

impixelinfo

figure('Name','A3{1}.BMd_Img_o_part1')

edges= linspace(0,255,21);

histogram(image1, 'BinEdges',edges,'normalization','pdf')

thanks a lot

[Fatemeh Shomal Zadeh]

Please let me know if you need any more information. I do really need to solve my problem as I need to report tomorrow in my meeting.

Thanks so much

[Star Strider]

It might be interesting to have the ‘image1’ vector you refer to here:

histogram(image1, 'BinEdges',edges,'normalization','pdf')

although havng it would not change anything I already wrote, so I doubt providing it would be worth the effort to attach it.

[Fatemeh Shomal Zadeh]

Unfortunately, image1 is not a vector, it is an array. When I am converting it to the vector as listed below it show this figure(i have attached here) which is not the figure that I am looking for

A = im2double(image1(:));

edges= linspace(0,255,21);

[muHat,sgHat] = normfit(A);

pdfplot = normpdf(edges, muHat, sgHat);

figure('Name','A3{1}.BMd_Img_o_part1')

histogram(image1, 'BinEdges',edges,'normalization','pdf')

hold on

plot(edges, pdfplot, '-r', 'LineWidth',2)

hold off

[Star Strider]

If you convert your image to grayscale and use imhist you may get the histogram you want.

Note that the histogram function will take matrix arguments as well as vector arguments, although it will convert a matrix to a single-column vector for analysis.

I am lost as to what you are doing.

[Fatemeh Shomal Zadeh]

I am sorry, maybe I was not clear enough in explaining my question.

Let me explain from the beginning. I have plotted this histogram (see attachment please) and now want to have a gaussian plot which has overlab with the histogram.

Data already are in a grayscale.

imshow(Data_SWS_BMd_A3{1}.BMd_Img_o)

impixelinfo

xmin = 5;

xmax = 314;

ymin = 136;

ymax = 238;

imshow(Data_SWS_BMd_A3{1}.BMd_Img_o(ymin:ymax,xmin:xmax));

impixelinfo

image= Data_SWS_BMd_A3{1}.BMd_Img_o(ymin:ymax,xmin:xmax)

[n,m]=size(image)

id=fix(m/3)

image1=image(:,1:id);

image2=image(:,id+1:2*id);

image3=image(:,2*id+1:m);

figure

imshow(image1)

impixelinfo

figure

imshow(image2)

impixelinfo

figure

imshow(image3)

impixelinfo

edges= linspace(0,255,21);

[muHat,sigmaHat] = normfit(image1);

pdfplot = normpdf(edges, muHat, sigmaHat);

figure('Name','A3{1}.BMd_Img_o_part1')

histogram(image1, 'BinEdges',edges,'normalization','pdf')

hold on

plot(edges, pdfplot, '-r', 'LineWidth',2)

hold off

I am receiving error like this which I do not know what it is;

Error in normfit

xc = x - repmat(muhat,[n 1]);

And if I convert the image1 to the vector double it shows the graph which is not right.

I know it is becoming complicated and thanks so much for your patience. If still it is not clear I do appologize for that.

Kind regards

[Star Strider]

I do not understand what threw the error.

Please use the code I already wrote to plot the normal distribution on the histogram.

[Fatemeh Shomal Zadeh]

Thank you so much , I do really appreciate you efforts.

[Star Strider]

As always, my pleasure!

Answer 2

[Fatemeh Shomal Zadeh]

this is a histogram figure that i have.