How to calculate peak value and duration of signal in Image

Hello,I have the following Image originalimage.png i want to calculate the peak value of the signal and also the duration of the each signal, as i attached the image parameterinstersted.png in which i describes which values i am currently determing

How can i do it in MATLAB

[Star Strider]

@Med Future — See my Answer. (posted about 12 hours earlier than your last Comment).

Answer 1

[Image Analyst]

Try this:

% Demo by Image Analyst
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 22;
markerSize = 40;

%--------------------------------------------------------------------------------------------------------
%    READ IN IMAGE
folder = [];
baseFileName = 'OriginalImage.png';
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
    % The file doesn't exist -- didn't find it there in that folder.
    % Check the entire search path (other folders) for the file by stripping off the folder.
    fullFileNameOnSearchPath = baseFileName; % No path this time.
    if ~exist(fullFileNameOnSearchPath, 'file')
        % Still didn't find it.  Alert user.
        errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
        uiwait(warndlg(errorMessage));
        return;
    end
end
grayImage = imread(fullFileName);
% Get the dimensions of the image.
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(grayImage)

%--------------------------------------------------------------------------------------------------------
% Display the image.
subplot(2, 2, 1);
imshow(grayImage);
impixelinfo;
axis('on', 'image');
title('Original Gray Scale Image', 'FontSize', fontSize, 'Interpreter', 'None');
if numberOfColorChannels > 1
    % It's not really gray scale like we expected - it's color.
    fprintf('It is not really gray scale like we expected - it is color\n');
    % Extract the blue channel.
    grayImage = grayImage(:, :, 3);
end
% Update the dimensions of the image.
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(grayImage)
% Maximize window.
g = gcf;
g.WindowState = 'maximized';
drawnow;

%--------------------------------------------------------------------------------------------------------
% Threshold the image to get the bright blobs.
mask = grayImage > 0;
subplot(2, 2, 2);
imshow(mask)
title('Binary Mask Image', 'FontSize', fontSize, 'Interpreter', 'None');

%--------------------------------------------------------------------------------------------------------
% Get the top row for each column
topRows = nan(1, columns);
for col = 1 : columns
    thisColumn = mask(:, col);
    % Find top row
    t = find(thisColumn, 1, 'first');
    if ~isempty(t)
        topRows(col) = t;
    end
end

% Now overlay it onto the image
subplot(2, 2, 3);
imshow(grayImage)
hold on;
plot(topRows, 'r-', 'LineWidth', 4)
title('Top Row in Red', 'FontSize', fontSize, 'Interpreter', 'None');
% Invert it and plot as a line plot
topRows = rows - topRows;
subplot(2, 2, 4);
plot(topRows, 'b-', 'LineWidth', 3);
grid on;
title('Line Plot', 'FontSize', fontSize, 'Interpreter', 'None');

[Med Future]

@Image Analyst Thanks for this, but this does not gives any value of each wave. and maximum height of the signal

[Image Analyst]

@Med Future I thought you'd be able to figure that out yourself once I got the signal vector.

Please invest 2 hours of your time here:

MATLAB Academy - Free 2 hour training

For one thing I didn't measure the width because you didn't specify the level at which you'd want the width for. You need to specify that.

And do you want just peaks taller than a certain height, or any peaks no matter how short they are?

Finding the peaks and valleys can be done with findpeaks if you have the Signal Processing Toolbox.

[peakValues, indexesOfPeaks] = findpeaks(topRows); % The indexex are column numbers in the image
[valleyValues, indexesOfValleys] = findpeaks(-topRows);
valleyValues = -valleyValues;

Now that you know peaks and valleys you can subtract indexes to get the distance in columns between the valleys.

[Med Future]

@Image Analyst

I want the width of two largest peak and the distance between two peaks, as i attached the image below.

Can you please help me

[Image Analyst]

Sure, if you let me. You didn't answer the questions I had in my last comment. Why didn't you???

[Med Future]

@Image Analyst I want the width peaks which have largest ,for example in the above picture i have only two largest peak, will ignore the smaller ones.

[Image Analyst]

Not sure why you're not doing what I said:

% Demo by Image Analyst
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 22;
markerSize = 40;
%--------------------------------------------------------------------------------------------------------
%    READ IN IMAGE
folder = [];
baseFileName = 'OriginalImage.png';
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
    % The file doesn't exist -- didn't find it there in that folder.
    % Check the entire search path (other folders) for the file by stripping off the folder.
    fullFileNameOnSearchPath = baseFileName; % No path this time.
    if ~exist(fullFileNameOnSearchPath, 'file')
        % Still didn't find it.  Alert user.
        errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
        uiwait(warndlg(errorMessage));
        return;
    end
end
grayImage = imread(fullFileName);
% Get the dimensions of the image.
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(grayImage)
%--------------------------------------------------------------------------------------------------------
% Display the image.
subplot(2, 2, 1);
imshow(grayImage);
impixelinfo;
axis('on', 'image');
title('Original Gray Scale Image', 'FontSize', fontSize, 'Interpreter', 'None');
if numberOfColorChannels > 1
    % It's not really gray scale like we expected - it's color.
    fprintf('It is not really gray scale like we expected - it is color\n');
    % Extract the blue channel.
    grayImage = grayImage(:, :, 3);
end
% Update the dimensions of the image.
% numberOfColorChannels should be = 1 for a gray scale image, and 3 for an RGB color image.
[rows, columns, numberOfColorChannels] = size(grayImage)
% Maximize window.
g = gcf;
g.WindowState = 'maximized';
drawnow;
%--------------------------------------------------------------------------------------------------------
% Threshold the image to get the bright blobs.
mask = grayImage > 0;
subplot(2, 2, 2);
imshow(mask)
title('Binary Mask Image', 'FontSize', fontSize, 'Interpreter', 'None');
%--------------------------------------------------------------------------------------------------------
% Get the top row for each column
topRows = nan(1, columns);
for col = 1 : columns
    thisColumn = mask(:, col);
    % Find top row
    t = find(thisColumn, 1, 'first');
    if ~isempty(t)
        topRows(col) = t;
    end
end
% Now overlay it onto the image
subplot(2, 2, 3);
imshow(grayImage)
hold on;
plot(topRows, 'r-', 'LineWidth', 4)
title('Top Row in Red', 'FontSize', fontSize, 'Interpreter', 'None');
% Invert it and plot as a line plot
topRows = rows - topRows;
subplot(2, 2, 4);
plot(topRows, 'b-', 'LineWidth', 3);
grid on;
title('Line Plot', 'FontSize', fontSize, 'Interpreter', 'None');
% Find peak heights
[peakValues, indexesOfPeaks] = findpeaks(topRows) % The indexex are column numbers in the image
[valleyValues, indexesOfValleys] = findpeaks(-topRows)
valleyValues = -valleyValues
% Find the separation of peaks
peakSeparation = indexesOfPeaks(4) - indexesOfPeaks(2)
% Find width of peak 1
peakWidth1 = indexesOfValleys(2) - indexesOfValleys(1)
% Find width of peak 2
peakWidth2 = indexesOfValleys(4) - indexesOfValleys(3)

If you do, you'll find:

peakSeparation =

321

peakWidth1 =

240

peakWidth2 =

239

[Med Future]

@Image Analyst i dont understand What you say?

[Med Future]

@Image Analyst you are using index value manually. But i want it to be automatically. if image change it does not work.

[Image Analyst]

I added in the findpeaks() code (in my last comment) that I suggested you do, and then computed the distances you wanted. It was trivial. This is what you get.

peakSeparation =

321

peakWidth1 =

240

peakWidth2 =

239

Distances are in pixels. Are those distances not what you wanted? If not, what do you want?

[Med Future]

@Image Analyst yes this distance i want but i am saying about the command

Should i put this in manually? is there is any other way? to get this automatically

peakSeparation = indexesOfPeaks(4) - indexesOfPeaks(2)
peakWidth1 = indexesOfValleys(2) - indexesOfValleys(1)
% Find width of peak 2
peakWidth2 = indexesOfValleys(4) - indexesOfValleys(3)

[Med Future]

@Image Analyst as you can see the following codes does not work it gives false values.

peakSeparation = indexesOfPeaks(4) - indexesOfPeaks(2)
peakWidth1 = indexesOfValleys(2) - indexesOfValleys(1)
% Find width of peak 2
peakWidth2 = indexesOfValleys(4) - indexesOfValleys(3)

[Image Analyst]

Not sure of your definition of "manually" versus "automatically". The code is in there to make the measurements and the results seem correct to me. I don't know why you said it's wrong, and you didn't even give me what you think the correct values were. But either way I'm sure you have enough skills now to do the modifications. I'm traveling over the next week and can't spend a lot of time on MATLAB Answers.

[Med Future]

@Image Analyst Let me explain it to you, when i give the next image i have attached above.

The code you share gives the different values on that as compared to actual output value.

Answer 2

[KALYAN ACHARJYA]

There are many threds sillimar quaestions, calculate the FWHM and time duration of the signal, please refer

https://in.mathworks.com/matlabcentral/answers/741247-how-to-find-fwhm-for-each-peak#answer_620307

https://in.mathworks.com/matlabcentral/answers/1697615-full-width-half-max#answer_945665

https://in.mathworks.com/matlabcentral/answers/1441689-i-am-trying-to-find-the-full-width-at-half-max-value-and-plot-the-waveform-with-markers?s_tid=srchtitle

Hope it Helps!

[Med Future]

@KALYAN ACHARJYA It does not help because i have image. and these were implemented on Signals

[KALYAN ACHARJYA]

Is that: Then getting the maximum distance white pixels from the max row number (same column case) consider as highest peak- Elcudian distance (But it is not the correct technical way to get the peak)?

Answer 3

[Star Strider]

The indexing in the image makes this a challenge, since they reflect back and in any event are non-monotonic, eliminating a straightforward solution and requiring a more intensive approach. The two peaks appear to be mirror images of each other, so I only analysed the left one here.

The Full-Width-Half-Maximum Value is 102.5 index units.

Try this —

Img = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/1169303/OriginalImage.png');

CC = bwconncomp(Img,8);
PIL = CC.PixelIdxList;

[yc,xc] = cellfun(@(x)ind2sub(CC.ImageSize,x), PIL, 'Unif',0);
xy = [cell2mat(xc.') cell2mat(yc.')];
xy(:,2) = max(xy(:,2)) - xy(:,2);

xyh = xy(1:fix(size(xy,1)/2),:);

figure
plot(1:size(xyh,1), xyh(:,1), 'DisplayName','Row Index')
hold on
plot(1:size(xyh,1), xyh(:,2), 'DisplayName','Column Index')
hold off
grid
legend('Location','best')

figure
plot((250:350), xyh(250:350,1), 'DisplayName','Row Index')
hold on
plot((250:350), xyh(250:350,2), 'DisplayName','Column Index')
hold off
grid
title('Original Vectors')
legend('Location','best')

figure
plot(xyh(:,1), xyh(:,2))
grid
xlabel('Row Index')
ylabel('Column Index)')

[maxv,ixh] = max(xyh(:,2));

xv1 = 1:ixh;                                                            % Lower Section Index Vector
xv2 = ixh+1:size(xyh,1);                                                % Upper Section Index Vector

figure
plot(xyh(xv1,1), xyh(xv1,2), 'DisplayName','Lower Section')
hold on
plot(xyh(xv2,1), xyh(xv2,2), 'DisplayName','Upper Section')
hold off
grid
xlabel('Row Index')
ylabel('Column Index)')
legend('Location','best')

Lv1 = diff([0; xyh(xv1,2)]) >= 1;                                       % Select Indices To Elliminate Discontinuities
Lv2 = diff([0; xyh(xv2,2)]) <= 1;                                       % Select Indices To Elliminate Discontinuities

figure
plot(xv1(Lv1(250:350)), xyh(xv1(Lv1(250:350)),1), 'DisplayName','Selected Row Index')
hold on
plot(xv1(Lv1(250:350)), xyh(xv1(Lv1(250:350)),2), 'DisplayName','Selected Column Index')
hold off
grid
title('Selected-Element Vectors')
xlabel('Absolute Index')
legend('Location','best')

ixv1 = xv1(Lv1);                                                        % Selected Indices
ixv2 = xv2(Lv2);                                                        % Selected Indices

yq = max(xyh(:,2))/2

yq = 499.5000

ix1 = find(diff(sign(xyh(ixv1,2)-yq)))+[0 1];                           % Approximate Index
xq(1) = interp1(xyh(ixv1(ix1),2), xyh(ixv1(ix1),1), yq);                % Interpolate

ix2 = find(diff(sign(xyh(ixv2,2)-yq)))+[0 1];                           % Approximate Index
xq(2) = interp1(xyh(ixv2(ix2),2), xyh(ixv2(ix2),1), yq);                % Interpolate

FWHM = xq(2)- xq(1)                                                     % FWHM: Desired Result (Units: Index)

FWHM = 102.5000

figure
plot(xyh(ixv1,1), xyh(ixv1,2))
hold on
plot(xyh(ixv2,1), xyh(ixv2,2))
hold off
grid
xlabel('Row Index')
ylabel('Column Index)')

.

[Med Future]

@Star Strider What is 102.500? as we can see the signal start from 250 and end at 425.

and it does not cater the maximum peak of the signal.

[Star Strider]

The ‘102.5’ value is the full-width-half-maximum (FWHM) value of the peak. (Because of the nature of the data, it is extremely difficult to draw it on the plot.) This is typically what is referred to as the ‘width’ of a peak Since you didn’t definie ‘width’, I used that value, since FWHM is the most commonly-used metric.

It may be difficult to determine with any accuracy where a peak begins and ends, as it is here as well, if you look at the data. There are several consecutive zero values at both minima near 250 and 425. Does the peak begin at the position of the first zero or the last in the blue portion of the curve, and similarly does it end at the first or last zero in the red pottion of the curve? The FWHM value is much easier to determine.

It calculates the maximum as ‘999’, the result of the max function, in order to get the ‘yq’ (half-maximum) value. It just doesn’t specifically report it.