How to extract the value pixel values from an image or masked image?

I need the values of the pixels from an image. I need values for each pixel separately.

[Zainab Afreen]

Hello I also have the same problem. Can anyone help me how do I detect pixels in an image automatically. Please reply

[Image Analyst]

This is so vague, just like the original question. If you have read in your image with imread() or gotten it some other way, like from a video camera, then you have your pixels already and you just need to reference them like the Answers below. Please explain in detail why the answers below do not give you any pixel values.

[DGM]

Oh that's easy. Say you have an image called myimage. Then to detect pixels:

haspixels = ~isempty(myimage);

You're welcome.

Answer 1

[Image Analyst]

Not sure what you mean. The image itself is a collection of pixel values. To get the pixel value at one particular (row, column) location, you can just specify the index:

grayLevel = grayImage(row, column);

or you can use impixel():

rgbColor = impixel(rgbImage, column, row);

[mohd akmal masud]

Hi Image Analyst, if i want list of pixel value that i have some point (ractangle)

let say i want list of pixel value from 58:71 58:74 85:75 90:60 71:71 (rectangle area), how to write the code?

this is my old code:

a = dicomread('I10.dcm' );
mask = (a >= 800) & (a <= 850);
for sliceIndex = 1 : size(a, 3)
    % Get 2-D mask at this slice level.
    thisMask = mask(:,:,sliceIndex);
    % Extract rows and columns for this slice only, not all of them.
    [rows, columns] = find(thisMask); 
    % Etc.  Now add these rows and columns to a master list of them over all slices.  (You do this)...
end

[Image Analyst]

Those don't look like any points on a rectangle (not the corner coordinates).

Assuming those are (row, column) and not (x, y), you could do

subImage = a(row1:row2, col1:col2);

but I can't tell what the first and last row and first and last column are from your numbers.

[mohd akmal masud]

It's work, but the result is the location of pixelm. How extract the value of pixel that point?

      subImage = a(75:79, 61:61)

[mohd akmal masud]

Hi Image Analyst.

I have been using your code for free hand masking. But how do I extract the list of pixel values from within the masked region?

[Image Analyst]

To get the value within a mask, for a grayscale image you can do

pixelValues = grayImage(mask);

For an RGB image I think you should split it into the 3 separate color channels then get them for each color channel.

To get the gray scale or RGB values of all the pixels in an image (not just the masked area though you could modify it to call fprintf only within the mask if you want), see the attached script which lets you read in an image and write out all the values to a CSV text file.

[Claudio Ignacio Fernandez]

Hello @Image Analyst

Today I used ginput over my image to get the x y coordinates. However, instead getting for example an x y values [40 80] I'm getting [40.12 80.07]. Why this is happening? any lead?

Anyway when using the weird X Y coordinates with the command impixel I get pixel values, however How can I know if I'm getting the value of the pixel I really want??

[Image Analyst]

That's the way ginput works - it gives you floating point values. You need to round to get array indexes of row and column

[x, y] = ginput(1);
row = round(y);
column = round(x);

Remember that row is y and column is x so don't make the mistake of saying yourImage(x, y) to reference pixels -- it's yourImage(y, x) which is yourImage(row, column).

pixelValue = yourImage(row, column); % If it's a gray scale image.
pixelValue = yourImage(row, column, :); % If it's an RGB Color image. pixelValue will be a 1x3 vector of [r,g,b] values.

Answer 2

[Ben11]

You can get the histogram of pixel values using imhist.

For example:

1) Grayscale image

Image = imread('coins.png');
[count,x] = imhist(Image);

2) RGB image

Image = imread('peppers.png');
[count,x] = imhist(Image(:,:,1)); % select one of 3 channels

or use rgb2gray:

   Image = imread('peppers.png');
   [count,x] = imhist(rgb2gray(Image));

[snehal jaipurkar]

this was helpful to me.after finding the count of each pixel value seperately, i want to add the counts of a range of pixel values and get the total, how to write its code in matlab?

[Image Analyst]

try this:

index1 = find(x == grayLevel1);
index2 = find(x == grayLevel2);
numPixels = sum(count(index1:index2));

[snehal jaipurkar]

thanks sir:) it worked :) i want to run this same program on suppose 50 images stored in a folder and based on the total count of pixels found as such in a certain range of graylevels is more than the total count of pixels found for other range of graylevels, i want to categorize that image into some category.how can i do dis sir???

[Image Analyst]

In the loop use an if statement to categorize:

if numPixels > threshold1
    % Do something
elseif numPixels > threshold2
    % Do something else
elseif numPixels > threshold3
    % Do a third thing
end

and so on.

[snehal jaipurkar]

and sir how to run this program on 50 images stored in a folder?

[Image Analyst]

Code samples are in the FAQ

Answer 3

[yonatan gerufi]

Hi Dhanya,

you can access to a specific pixel by typing : figure_name(x_pos,y_pos) .

In the MATLAB workspace, most images are represented as two-dimensional arrays (matrices), in which each element of the matrix corresponds to a single pixel in the displayed image. (from Matlab documentation )

This matrix can be represented in several types as double, uint8, uint16. It can also be RGB, intensity, or indexed types.

I highly recommend reading to understand the differences.

see image types , and working with images.

[Image Analyst]

That's not exactly correct. The correct indexing is given in my answer. Mixing up x,y with row,column is a very common error. To fix your reference you'd do figure_name(y_pos,x_pos) (reversing the indexes as you had them) where figure_name is the the name of the image variable rather than the figure. In short, first index = row = y (not x), and second index = column = x (not y).

[Madhava Teja Munagala]

For example I have rice image, How to find no of pixels in image,,, How to write code for that

[Image Analyst]

Multiply the rows by the columns.

[rows, columns, numberOfColorChannels] = size(riceImage);
numPixelsInImage = rows * columns;

This gives you the total number of pixels in the entire image, rice and background included.

[Gustavo Liñán Cembrano]

If your image is single channel (grayscale), then you simply do:

NumPixelsInImage=length(riceImage(1:end));

Getting familiar with the (1:end) operation on images is useful sometimes.

[Guillaume]

Actually, getting familiar with numel which in all respects is much better than length would be better. Your code is simply:

numPixelsInImage = numel(riceImage);

which will be faster as it doens't ask matlab to reshape the image (a waste of time) just so you can know how many elements there are.

[Image Analyst]

numel() is a good way for grayscale images. For color images, numel doesn't give you the number of pixels since most people consider the number of lateral points pixels, so a single pixel would have 3 elements for a color image.

Not sure what was meant by reshaping the image, but size() does not reshape the image, nor does doing riceImage(:) or riceImage(1:end). I also thought it did until a Mathworks developer assured me that it does not reshape or cause any movement of values in RAM memory or allocating additional memory even though (:) gives a 1-D column vector.

[Guillaume]

My comment was addressed to Gustavo whose code is just a more convoluted

numPixelsInImage = numel(riceImage);

His length(array(1:end)) (which is the same as length(array(:)) by the way) reshapes the image into a vector just to know how many pixels there are. Indeed it (and numel) only works with grayscale images.

And, yes reshape does not reorder the pixels. However, it does use some cpu cycles for a completely unnecessary operation.

[Image Analyst]

I was told by a Mathworker that (:) does not reshape the array into a column vector. So passing it into size() or length() would not reshape it. He said that if you assign that to a NEW variable, that new variable will be of a columnar shape, but that is a new variable and there is no temporary variable that is created with a column shape nor is the original array reshaped into a column vector. Only a new variable would have that shape. Fine point though.

Answer 4

[Youssef Khmou]

Accessing a pixel is similar to retrieving element from matrix, here are two examples :

for gray scale image :

X=imread('circuit.tif');
X(10,60)

for multi channel image :

Y=imread('autumn.tif');
Y(10,60,1) %R
Y(10,60,2) %G

Answer 5

[snehal jaipurkar]

after finding the count of each pixel value seperately, i want to add the counts of a range of pixel values and get the total, how to write its code in matlab??

[SHILPA K K]

Use imhist()

[Image Analyst]

Use sum:

sumOfCounts = sum(counts(index1:index2));

where index1 and index2 define the "range of pixel values" that you want to sum over.

Answer 6

[Umar Awan]

how can i convert an 28*28 pixel image into 1*784 ? means in 1 row

[Image Analyst]

If grayImage is your 2-D image, do

rowImage = reshape(grayImage, 1, []);

to reshape the 2-D gray scale image into a 1-D row vector.

Answer 7

[Asad Alam]

How can i compare pixel value of an image

pixelvalue<300

And i want all the pixels whose values are above 300. can anyone help

[Gustavo Liñan]

% Since you're specifying a single scalar threshold, I assume your image is grayscale
% If you want a binary of the same size just use, assuming your image is already in memory
% in variable YourImage;
YourThreshold=300;
RESULT=YourImage<YourThreshold;

%If your image is in a File;
YourImage=imread(fullfile(YourImagePath,YourImageFileName));
RESULT=YourImage<YourThreshold;

%Now if you want this result as a vector of Ncols x Nrows elements, just do
RESULT_VECTOR=RESULT(1:end); 
%The indexes for these pixels are simply
MY_PIX_IDX=find(RESULT_VECTOR)

%Now if you want to extract the pixel values satisfying the condition

YOUR_PIXELS_BELOW_THRES=YourImage(MY_PIX_IDX);

The question is indeed a bit vague to give a completely operative solution, hope these hints help.

[Image Analyst]

This is not an answer to @Dhanya and should have been it's own question. Anyway...

To get a binary image "map" of where those pixels are, you can do

binaryImage = yourImage < 300;

Now these might be some irregularly-shaped "blobs" as we call them in image processing. To extract those pixels in the blobs into one giant list (vector), you can do

listOfPixelValues = yourImage(binaryImage);

If you want the values of each blob separate from all other blobs, you need to call regionprops

props = regionprops(binaryImage, 'PixelValues');

props is a structure array. You can also ask for the data to come back in a table instead of a structure array if you want.