I have to process a very large image ( say 10 MB image file or even more).I have to remove artifacts and dead pixels in MATLAB
I have read about Block Processing of Large Images, but have no idea how to apply it to a 16 bit image.
I am referring to removal of pixels which have highest value into the average value of surrounding pixel .my code is not working on my image which is 80 MB of size
numIterations = 30;
avgPrecisionSize = 16; % smaller is better, but takes longer
% Read in the image grayscale:
originalImage = double(rgb2gray(imread('C:\Documents and Settings\admin\Desktop\TM\image5.tif')));
% get the bad pixels where = 0 and dilate to make sure they get everything:
badPixels = (originalImage == 0);
badPixels = imdilate(badPixels, ones(12));
%# Create a big gaussian and an averaging kernel to use:
G = fspecial('gaussian',[1 1]*100,50);
H = fspecial('average', [1,1]*avgPrecisionSize);
%# User a big filter to get started:
newImage = imfilter(originalImage,G,'same');
newImage(~badPixels) = originalImage(~badPixels);
% Now average to
for count = 1:numIterations
newImage = imfilter(newImage, H, 'same');
newImage(~badPixels) = originalImage(~badPixels);
end
%% Plot the results
figure(123);
clf;
% Display the mask:
subplot(1,2,1);
imagesc(badPixels);
axis image
title('Region Of the Bad Pixels');
% Display the result:
subplot(1,2,2);
imagesc(newImage);
axis image
set(gca,'clim', [0 255])
title('Infilled Image');
colormap gray
newImage2 = roifill(originalImage, badPixels);
figure(44);
clf;
imagesc(newImage2);
colormap gray
You are doing a few things which are obvious problems (but it might depend specifically on how far you can get into the code before you run out of memory)
1) You are immediately converting the whole image to double
2) You are identifying certain pixels which you want to replace, but passing the whole image to imfilter and then throwing away (presumably) most of the output:
newImage = imfilter(originalImage,G,'same'); % filter across the entire image
newImage(~badPixels) = originalImage(~badPixels); % replace all the good pixels!
Without converting to double, why not first check where the bad pixels are, do your processing on subregions of the appropriate size around those pixels (the subregions can be converted to double and back), and then reassemble the image at the end?
blockproc may work if you can write your filtering option as a function which takes in an image area and returns the correct area - you'll have to use the border_size option appropriately and make sure your function just returns the original image without bothering to do any filtering if it hits a block with no bad pixels in. You can even have it write out to file as well.
Related
I'm new to image processing and started using MATLAB for Astrophotography processing. I'm trying to process 10 corrupted images (same image but mixed with different noise) of the planet Saturn using MATLAB. I learned that by stacking the 10 images together leads to a noise reduced picture with high PSNR and tried the below coding to make it work.
But the output looks like an unclear saturated image with no noise reduction.
Can you please look at the code below and show me where I went wrong?
%% We are going to stack the 10 corrupted images and finally calculate the PSNR SSIM
clearvars;% Clear all the variables
close all;
load('planetdata.mat'); %to load the corrupted Image set (4-D uint8)
Clean = imread('Clean Image of Saturn.jpg');%Clean Image of Saturn.600x800x3 uint8
planet1(: , :, :) = planetdata(1, :, :, :);%One corrupted Image as reff
% Set the number of images to stack is 10
stack_number = 10;
% Lets use Clean image as reference of dimensions required
im_x = size(Clean, 1);
im_y = size(Clean, 2);
im_z = size(Clean, 3);
% Lets Generate a blank image for image stacking
resultIM = uint8(zeros(im_x, im_y, im_z));
% Iterate through the images to stack
for i = 1:1:stack_number
% Read in the target object image
CorruptIM(: , :, :) = planetdata(i, :, :, :);
% Perform image stacking using the target object image
resultIM = resultIM + CorruptIM;
end
% resultIM = resultIM / stack_number;
%% Lets Display Results
workspace; % to Make sure the work space panel is showing.
fontSize = 15;
figure;
subplot(1, 3, 1);
imshow(Clean);
title('Clean Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'Position', get(0,'Screensize'));
% Give a name to the title bar.
set(gcf,'name','Stacking','numbertitle','off')
% Display one corrupt image as reference
subplot(1, 3, 2);
imshow(planet1);
title('Corrupt Image 1 : Ref', 'FontSize', fontSize);
% Display Stacked image
subplot(1, 3, 3);
imshow(resultIM);
title('Stacked Image', 'FontSize', fontSize);
%% PSNR AND SSIM Calculation
%Lets Find PSNR for For Resultant Image
[row,col] = size(Clean);
size_host = row*col;
o_double = double(Clean);
w_double = double(resultIM);
s=0;
for j = 1:size_host % the size of the original image
s = s+(w_double(j) - o_double(j))^2 ;
end
mes =s/size_host;
psnr =10*log10((255)^2/mes);
fprintf('The PSNR value for Stacked Image is %0.4f.\n',psnr);
%Lets Find SSIM for resultant Image
[ssimval, ssimmap] = ssim(uint8(resultIM),Clean);
fprintf('The SSIM value for Stacked Image is %0.4f.\n',ssimval);
I think this one statement pretty much says it all (emphasis mine):
But the output looks like an unclear saturated image with no noise reduction.
It looks like your images are in fact saturating at the upper limit for a uint8 variable, which is the data type for your result image resultIM and your data matrix planetdata. As you keep adding images, the pixel values saturate at the maximum value of 255 for unsigned 8-bit integer types.
What you'll need to do is convert to a larger data type for your intermediate calculations, such as a larger integer type or a floating point type. Then, once your calculations are complete (e.g. dividing the sum by the image stack size to get an average image), you can scale and/or round your data as needed and convert it back to a uint8 type.
I am trying to "translate" what's mentioned in Gonzalez and Woods (2nd Edition) about the Laplacian filter.
I've read in the image and created the filter. However, when I try to display the result (by subtraction, since the center element in -ve), I don't get the image as in the textbook.
I think the main reason is the "scaling". However, I'm not sure how exactly to do that. From what I understand, some online resources say that the scaling is just so that the values are between 0-255. From my code, I see that the values are already within that range.
I would really appreciate any pointers.
Below is the original image I used:
Below is my code, and the resultant sharpened image.
Thanks!
clc;
close all;
a = rgb2gray(imread('e:\moon.png'));
lap = [1 1 1; 1 -8 1; 1 1 1];
resp = uint8(filter2(lap, a, 'same'));
sharpened = imsubtract(a, resp);
figure;
subplot(1,3,1);imshow(a); title('Original image');
subplot(1,3,2);imshow(resp); title('Laplacian filtered image');
subplot(1,3,3);imshow(sharpened); title('Sharpened image');
I have a few tips for you:
This is just a little thing but filter2 performs correlation. You actually need to perform convolution, which rotates the kernel by 180 degrees before performing the weighted sum between neighbourhoods of pixels and the kernel. However because the kernel is symmetric, convolution and correlation perform the same thing in this case.
I would recommend you use imfilter to facilitate the filtering as you are using methods from the Image Processing Toolbox already. It's faster than filter2 or conv2 and takes advantage of the Intel Integrated Performance Primitives.
I highly recommend you do everything in double precision first, then convert back to uint8 when you're done. Use im2double to convert your image (most likely uint8) to double precision. When performing sharpening, this maintains precision and prematurely casting to uint8 then performing the subtraction will give you unintended side effects. uint8 will cap results that are negative or beyond 255 and this may also be a reason why you're not getting the right results. Therefore, convert the image to double, filter the image, sharpen the result by subtracting the image with the filtered result (via the Laplacian) and then convert back to uint8 by im2uint8.
You've also provided a link to the pipeline that you're trying to imitate: http://www.idlcoyote.com/ip_tips/sharpen.html
The differences between your code and the link are:
The kernel has a positive centre. Therefore the 1s are negative while the centre is +8 and you'll have to add the filtered result to the original image.
In the link, they normalize the filtered response so that the minimum is 0 and the maximum is 1.
Once you add the filtered response onto the original image, you also normalize this result so that the minimum is 0 and the maximum is 1.
You perform a linear contrast enhancement so that intensity 60 becomes the new minimum and intensity 200 becomes the new maximum. You can use imadjust to do this. The function takes in an image as well as two arrays - The first array is the input minimum and maximum intensity and the second array is where the minimum and maximum should map to. As such, I'd like to map the input intensity 60 to the output intensity 0 and the input intensity 200 to the output intensity 255. Make sure the intensities specified are between 0 and 1 though so you'll have to divide each quantity by 255 as stated in the documentation.
As such:
clc;
close all;
a = im2double(imread('moon.png')); %// Read in your image
lap = [-1 -1 -1; -1 8 -1; -1 -1 -1]; %// Change - Centre is now positive
resp = imfilter(a, lap, 'conv'); %// Change
%// Change - Normalize the response image
minR = min(resp(:));
maxR = max(resp(:));
resp = (resp - minR) / (maxR - minR);
%// Change - Adding to original image now
sharpened = a + resp;
%// Change - Normalize the sharpened result
minA = min(sharpened(:));
maxA = max(sharpened(:));
sharpened = (sharpened - minA) / (maxA - minA);
%// Change - Perform linear contrast enhancement
sharpened = imadjust(sharpened, [60/255 200/255], [0 1]);
figure;
subplot(1,3,1);imshow(a); title('Original image');
subplot(1,3,2);imshow(resp); title('Laplacian filtered image');
subplot(1,3,3);imshow(sharpened); title('Sharpened image');
I get this figure now... which seems to agree with the figures seen in the link:
I have two sets of images of different size for each set. The first set is images of 400x400 pixels with real picture objects.
The second set is 319x319, with image silhouettes of different scale than the real picture objects.
What I want to achieve, is basically to have the silhouettes replaced by the real picture objects (i.e. beaver) of the first set. So the end result will be 319x319 resolution images with real picture objects. Here is an example:
The first set images cannot simply be resized to 319x319, since the beaver will not match the silhouette. There are about 100 images with different "beaver size to beaver's silhouette size" relationships. Is there a way to automate this procedure?
So far, I've tried #cxw suggestion up to step 2. Here is the code of EllipseDirectFit I used. And here is my code to plot the images with the ellipse fits. I don't know how to proceed to steps 3-5.. I think from EllipseDirectFit function -> 2*abs(A(1)) should be the ellipsi's major axes. (NOTE: 'a1.bmp' is the real image and 'b1.bmp' is the silhouette).
In case anyone else has the same problem as me, I post the code that solved my problem. I actually followed cxw's suggestion and fitted an ellipse for both real and silhouette pictures, then resized the real picture based on the ratio of the silhouette-ellipse's major axis to the real-ellipse major axis. This made the image object match in size the silhouette image object (i.e. the beaver). Then I either cropped, or added border pixels to match the resolution I needed (i.e. 319x319).
% fetching the images
realList = getAllFiles('./real_images'); % getAllFiles => StackOverflow function
silhList = getAllFiles('./silhouettes');
for qq = 1:numel(realList)
% Name of the file to save
str = realList{qq}(15:end);
a = imread(realList{qq}); % assign real image
background_Ra = a(1,1,1); % getting the background colors
background_Ga = a(1,1,2);
background_Ba = a(1,1,3);
% finding the points (x,y) to pass to fit_ellipse
[x1,y1]=find(a(:,:,1)~=background_Ra | a(:,:,2)~=background_Ga | a(:,:,3)~=background_Ba);
% fitting an ellipse to these points
z1 = fit_ellipse(x1,y1); % Mathworks file exchange function
b = imread(silhList{qq}); % assign silhouette image
background_R2b = b(1,1,1); % getting the background colors
background_G2b = b(1,1,2);
background_B2b = b(1,1,3);
% finding the points (x,y) to pass to fit_ellipse
[x2,y2]=find(b(:,:,1)~=background_R2b & b(:,:,2)~=background_G2b & b(:,:,3)~=background_B2b);
% fitting an ellipse to these points
z2 = fit_ellipse(x2,y2);
% ratio of silhouette's ellipse major axis to real image's ellipse
% major axis
ellaxratio = z2.long_axis/z1.long_axis;
% resizing based on ellaxratio, so that the real image object size will
% now fit the silhouette's image object size
c = imresize(a,ellaxratio); c = rgb2gray(c);
bordercolor = c(end,end);
% if the resulting image is smaller, add pixels around it until they
% match with the silhouette image resolution
if size(c) < 319
while size(c) < 319
% 'addborder' is a Mathworks file exchange function
c = addborder(c(:,:,1),1, bordercolor ,'outer');
end
% if the resulting image is larger, crop pixels until they match
else size(c) > 319
while size(c) > 319
c = c(2:end-1,2:end-1);
end
end
% in a few cases, the resulting resolution is 318x318, instead of
% 319x319, so a small adjustment won't hurt.
if size(c) ~= 319
c = imresize(c,[319 319]);
end
% saving..
imwrite(c,['./good_fits/' str '.bmp'])
end
I don't have code for this, but here's how I would proceed, just off-hand. There's almost certainly a better way :) .
For each of the real image and the silhouette image:
Get the X, Y coordinates of the pixels that aren't the background. Edit Example tested in Octave:
background_R = img(1,1,1)
background_G = img(1,1,2)
background_B = img(1,1,3)
[xs,ys]=find(img(:,:,1)~=background_R | img(:,:,2)~=background_G | img(:,:,3)~=background_B)
The logical OR is because the image can differ from the background in any color component.
Fit an ellipse to the X, Y coordinate pairs you found. E.g., use this routine from File Exchange. (Actually, I suppose you could use a circle fit or any other shape fit you wanted, as long as size and position are the only differences between the non-background portions of the images.)
Now you have ellipse parameters for the real image and the silhouette image. Assuming the aspect ratios are the same, those ellipses should differ only in center and scale.
Resize the real image (imresize) based on the ratio of silhouette ellipse major axis length to real image ellipse major axis length. Now they should be the same size.
Find the centers. Using the above fit routine,
A=EllipseDirectFit(...)
% switch to Mathworld notation from http://mathworld.wolfram.com/Ellipse.html
ma=A(1); mb=A(2)/2; mc=A(3); md=A(4)/2; mf=A(5)/2; mg=A(6);
center_x = (mc*md-mb*mf)/(mb**2-ma*mc)
center_y = (ma*mf-mb*md)/(mb**2-ma*mc)
Move the real image data in a 3-d matrix so that the ellipse centers
coincide. For example,
cx_silhouette = ... (as above, for the silhouette image)
cy_silhouette = ...
cx_real = ... (as above, for the *resized* real image)
cy_real = ...
shifted = zeros(size(silhouette_image)) % where we're going to put the real image
deltax = cx_silhouette - cx_real
deltay = cy_silhouette - cy_real
% if deltax==deltay==0, you're done with this step. If not:
portion = resized_real_image(max(deltay,0):319-abs(deltay), max(deltax,0):319-abs(deltax), :); % or something like that - grab the overlapping part of the resized real image
shifted(max(deltay,0):min(deltay+319,319), max(deltax,0):min(deltax+319,319), :) = portion; % or something like that - slide the portion of the resized real image in x and y. Now _shifted_ should line up with the silhouette image.
Using the background color (or the black silhouette — same difference) as a mask, copy pixels from the resized, moved real image into the silhouette image.
Hope this helps!
does anybody know how to crop an image with a sliding window in Matlab?
e.g. I have an image of 1000x500 pixels, I would like to crop from this image blocks of 50x50 pixels... Of course I have to handle uneven divisions, but it is not necessary to have block of the same size.
Some details that have helped me in the past related to (i) ways to divide an image while block processing and (ii) "uneven division", as mentioned by OP.
(i) Ways to divide/process an image:
1. Process non-overlapping blocks:
Using default parameter {'BorderSize',[0 0]}, this can be handled with blockproc as below.
Example for (i)-1: Note the blocked nature of the output. Here each non-overlapping block of size 32 x 32 is used to calculate the std2() and the output std2 value is used to fill that particular block. The input and output are of size 32 x 32.
fun = #(block_struct) std2(block_struct.data) * ones(size(block_struct.data));
I2 = blockproc('moon.tif',[32 32],fun);
figure; subplot(1, 2, 1);
imshow('moon.tif'); title('input');
subplot(1,2, 2)
imshow(I2,[]); title('output');
Input and Output Image:
(i)-2: Process overlapping blocks:
Using parameter {'BorderSize',[V H]}: V rows are added above and below the block and H columns are added to the left and right of the block. The block that is processed has (N + 2*V) rows and (M + 2*H) columns. Using default parameter {'TrimBorder',true}, the border of the output is trimmed to the original input block size of N rows and M columns.
Example for (i)-2: Below code using blockproc uses {'BorderSize',[15 15]} with [N M] = [1 1]. This is similar to filtering each pixel of the image with a custom kernel. So the input to the processing unit is a block of size (1 + 2*15) rows and (1 + 2*15) columns. And since {'TrimBorder',true} by default, the std2 of the 31 rows by 31 columns block is provided as output for each pixel. The output is of size 1 by 1 after trimming border. Consequently, note that this example output is 'non-blocked' in contrast to the previous example. This code takes much longer time to process all the pixels.
fun = #(block_struct) std2(block_struct.data) * ones(size(block_struct.data));
I2 = blockproc('moon.tif',[1 1],fun,'BorderSize',[15 15]);
figure; subplot(1, 2, 1);
imshow('moon.tif'); title('input');
subplot(1,2, 2)
imshow(I2,[]); title('output');
Input and Output Image:
(ii) "Uneven division":
1. Zero/replicate/symmetric padding:
Zero padding so that an integer multiple of the blocks (N rows by M cols sized) can cover the [image + bounding zeros] in the uneven dimension. This can be achieved by using the default parameter {'PadMethod', 0} along with {'PadPartialBlocks' , true} ( which is false by default ). If a bounding region of zeros causes a high discontinuity in values computed from the bounding blocks, {'PadMethod', 'replicate'} or {'PadMethod', 'symmetric'} can be used.
2. Assume an "Active Region" within the image for block processing
For the case of processing each pixel, as in case (i)-2, we could assuming a bounding region of floor(block_size/2) pixels on all sides along the periphery of the image that is used as "Dummy" region. The Active region for block processing is contained within the Dummy region.
Something similar is used in imaging sensors where Dummy Pixels located at the periphery of an imaging array of Active Pixels allow for an operation like the color interpolation of all active area pixels. As color interpolation usually needs a 5x5 pixel mask to interpolate the color values of a pixel a bounding Dummy periphery of 2 pixels can be used.
Assuming MATLAB indexing, the region ( floor(block_size/2) + 1 ) to ( Input_Image_Rows - floor(block_size)/2) ) Rows by ( floor(block_size/2) + 1 ) to ( Input_ImageCols - floor(block_size)/2) ) Columns is considered as Active region (assuming a square block of side, block_size) which undergoes block processing for each pixel as in (i)-2.
Assuming a square block size of 5 by 5, this is shown by the following:
block_size = 5;
buffer_size = floor(block_size/2);
for i = (buffer_size+1):(image_rows-buffer_size)
for j = (buffer_size+1):(image_cols-buffer_size)
... % block processing for each pixel Image(i,j)
end
end
Matlab ver: R2013a
I would first look into the function blockproc to see if it can do what you want.
If you're sure you want to manually crop the image into blocks, you can use this script. It both writes the cropped images to .png files and saves the cropped images in the pages of a 3D array. You can modify it as you need.
This script assumes the image in evenly divisible by the block size. If it isn't, you'll need to pad it with zeros.
[rowstmp,colstmp]= size(myImage);
block_height = 50;
block_width = 50;
blocks_per_row = rows/block_height;
blocks_per_col = cols/block_width;
number_of_blocks = blocks_per_row*blocks_per_col;
%// pad image with zeros if needed
if ~(mod(rowstmp-1,block_height)==0)
rows = ceil(rowstmp/block_height)*block_height;
end
if ~(mod(colstmp-1,block_width)==0)
cols = ceil(colstmp/block_width)*block_width;
end
Im = uint8(zeros(rows,cols));
Im(1:rowstmp,1:colstmp) = myImage;
%// make sure these image have type uint8 so they save properly
cropped_image = uint8(zeros(rows,cols));
img_stack = uint8(zeros(rows,cols,number_of_blocks));
%// loop over the image blocks
for i = 1:blocks_per_row
for j = 1:blocks_per_col
%// get the cropped image from the original image
idxI = 1+(i-1)*block_height:i*block_height;
idxJ = 1+(j-1)*block_width :j*block_width;
cropped_image(idxI,idxJ) = Im(idxI,idxJ);
%//imshow(cropped_image)
%// write the cropped image to the current folder
filename = sprintf('block_row%d_col%d.png',i,j);
imwrite(cropped_image,filename);
cropped_image(idxI,idxJ) = 0;
%// keep all the blocks in a 3D array if we want to use them later
img_stack(:,:,(i-1)*blocks_per_col+j);
end
end
I have 8 images and I want to show them in a scale-space format shown below. The original image height and width is 256. Then on right side of original image, at every level the size is reduced by 2. Like here, image height and width is 256. On the right side of original image, height and width is 128, 64, 32, 16, 8, 4 , 2.
I have all the images in their respective dimensions. I just want to know how do I arrange the images according the pattern shown below. Thanks in advance.
This looks like you are trying to build a scale-space and displaying the results to the user. This isn't a problem to do. Bear in mind that you will have to do this with for loops, as I don't see how you will be able to do this unless you copy and paste several lines of code. Actually, I'm going to use a while loop, and I'll tell you why soon.
In any case, you need to declare an output image that has as many rows as the original image, but the columns will be 1.5 times the original image to accommodate for the images on the right.
First, write code that places the original image on the left side, and the version that is half the size on the right. Once you do this, you write a for loop, use indexing to place the images in the right spots until you run out of scales, and you need to keep track of where the next image starts and next image ends. Bear in mind that the origin of where to write the next images after the first subsampled one will start at the column location of the original image, and the row is right where the previous one ends. As an example, let's use the cameraman.tif image that is exactly 256 x 256, but I will write code so that it fits any image resolution you want. When I subsample the image, I will use imresize in MATLAB to help with the resizing of the image, and I'll specify a sampling factor of 0.5 to denote the subsampling by 2. The reason why I would use a while loop instead, is because we can keep looping and resizing until one of the dimensions of the resized image is 1. When this is the case, there are no more scales to process, and so we can exit.
As such:
%// Read in image and get dimensions
im = imread('cameraman.tif');
[rows,cols] = size(im);
%// Declare output image
out = zeros(rows, round(1.5*cols), 'uint8');
out(:,1:cols) = im; %// Place original image to the left
%// Find first subsampled image, then place in output image
im_resize = imresize(im, 0.5, 'bilinear');
[rows_resize, cols_resize] = size(im_resize);
out(1:rows_resize,cols+1:cols+cols_resize) = im_resize;
%// Keep track of the next row we need to write at
rows_counter = rows_resize + 1;
%// For the rest of the scales...
while (true)
%// Resize the image
im_resize = imresize(im_resize, 0.5, 'bilinear');
%// Get the dimensions
[rows_resize, cols_resize] = size(im_resize);
%// Write to the output
out(rows_counter:rows_counter+rows_resize-1, cols+1:cols+cols_resize) = ...
im_resize;
%// Move row counter over for writing the next image
rows_counter = rows_counter + rows_resize;
%// If either dimension gives us 1, there are no more scales
%// to process, so exit.
if rows_resize == 1 || cols_resize == 1
break;
end
end
%// Show the image
figure;
imshow(out);
This is the image I get: