How I can know the size of a quantized image

Question

Osama Hussein 2016-2-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/268122-how-i-can-know-the-size-of-a-quantized-image

评论： Osama Hussein 2016-2-15

I designed a simple uniform quantizer, it is a part of project, I used 'leena.jpg' image the original size is about 65.8 KB, after running this code:

img = imread('leena.jpg');
figure(1)
imshow(img)
[m,n] = size(img);
M = 8; K = 256/M;
for i = 1:m
      for j = 1:n
          quant_img(i,j) = floor(img(i,j)/K) * K + (K/2);
      end
end
figure(2)
imshow(quant_img)

-----------------

It is clear the image quality reduced, but when I save the two figures (the original and the quantized) as jpg format, I got 14.4 KB for the original and 16.5 KB for the quantized image! When I increase the number of levels (M = 32) the size decrease to 14.8 KB! Why the size of the image is decreasing with increasing the number of levels!! It should be the reverse, right?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2016-2-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/268122-how-i-can-know-the-size-of-a-quantized-image#answer_209838

It's a moot question because you should not even be using jpg. Use PNG instead. It will be bigger but at least you won't have JPG artifacts because it's lossless compression, and pretty much the new standard that most people use.

7 个评论
显示 5更早的评论隐藏 5更早的评论

Image Analyst 2016-2-15

在 MATLAB Online 中打开

It doesn't. Run this code and tell me what you observe:

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 20;
% img = imread('leena.jpg');
img = imread('cameraman.tif');
subplot(1,2,1);
imshow(img)
axis on;
title('Original Image', 'FontSize', fontSize);
[rows, columns, numberOfColorChannels] = size(img)
% Set up figure properties:
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Get rid of tool bar and pulldown menus that are along top of figure.
set(gcf, 'Toolbar', 'none', 'Menu', 'none');
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')
message = sprintf('For quantized image\nRows = %d\nColumns = %d\nNumber of Color Channels = %d', rows, columns, numberOfColorChannels);
uiwait(helpdlg(message));
M = 8; 
K = 256/M;
for row = 1 : rows  
  for col = 1 : columns    
    quant_img(row,col) = floor(img(row,col)/K) * K + (K/2);    
  end  
end
subplot(1,2,2);
imshow(quant_img)
axis on;
[rows, columns, numberOfColorChannels] = size(quant_img)
title('Quantized Image', 'FontSize', fontSize);
message = sprintf('For quantized image\nRows = %d\nColumns = %d\nNumber of Color Channels = %d', rows, columns, numberOfColorChannels);
uiwait(helpdlg(message));

For the cameraman image it remains 256x256. Change it to your lena image and see what it says. It's not a color image is it? The numbers don't even make sense for a color image so I think there's something you're not sharing. There might be more code there that you've left out. Or you didn't clear variables and close figure windows so maybe something is still there from an earlier run where you had different values for the variables. Like maybe you didn't close the figures and put hold on, and then changed the image file to one with a different size.

Image Analyst 2016-2-15

You're probably saving the figure window itself, rather than just the image. The figure will include things like gray background, tick marks, annotations in the graphical overlay, etc.

Use imwrite() rather than saveas().

Osama Hussein 2016-2-15