Image compression huffman coding

%clearing all variableas and screen
clear all;
close all;
clc;

%Reading image
a=imread('jpeg-image-compression-1-638.JPG');
figure,imshow(a)

%converting an image to grayscale
I=rgb2gray(a);

%size of the image
[m,n]=size(I);
Totalcount=m*n;

%variables using to find the probability
cnt=1;
sigma=0;

%computing the cumulative probability.
for i=0:255
k=I==i;
count(cnt)=sum(k(:))

%pro array is having the probabilities
pro(cnt)=count(cnt)/Totalcount;
sigma=sigma+pro(cnt);
cumpro(cnt)=sigma;
cnt=cnt+1;
end;


%Symbols for an image
symbols = [0:255];

%Huffman code Dictionary
dict = huffmandict(symbols,pro);

%function which converts array to vector
vec_size = 1;
for p = 1:m
for q = 1:n
newvec(vec_size) = I(p,q);
vec_size = vec_size+1;
end
end

%Huffman Encodig
hcode = huffmanenco(newvec,dict);

%Huffman Decoding
dhsig1 = huffmandeco(hcode,dict);

%convertign dhsig1 double to dhsig uint8
dhsig = uint8(dhsig1);

%vector to array conversion
dec_row=sqrt(length(dhsig));
dec_col=dec_row;

%variables using to convert vector 2 array
arr_row = 1;
arr_col = 1;
vec_si = 1;

for x = 1:m
for y = 1:n
back(x,y)=dhsig(vec_si);
arr_col = arr_col+1;
vec_si = vec_si + 1;
end
arr_row = arr_row+1;
end


%converting image from grayscale to rgb
[deco, map] = gray2ind(back,256);
RGB = ind2rgb(deco,map);
imwrite(RGB,'decoded.JPG');

%end of the huffman coding

This code works fine but the image decoded is grey in colour. I cannot find anything wrong in it please help

[Bhaumik chaudhari]

hey , can you send me the code that is working for image processing huffman coding

email id:[e-mail]

[Walter Roberson]

1. Read the image.
2. reshape the image to be a vector.
3. Use histcounts or histc to count the number of occurances of each of the bytes
4. throw away any entries that have a count of 0 (but keep a list of what the original value is for each)
5. Create the huffman dictionary.
6. Use the huffman encoding function.
7. Store the image size and the huffman dictionary and the huffman encoding vector efficiently .Most people miss this step.

[nor saziana ariani sazali]

Hi, walter..can i know step for storing huffman dictionary after encoding using huffman?

[Walter Roberson]

https://en.wikipedia.org/wiki/Canonical_Huffman_code

[maha asiri]

did you find the anwer for this code

couse i want it

Answer 1

[Saherish Pathan]

clear all;
close all;
clc;
%Reading image
a=imread('jpeg-image-compression-1-638.JPG');
figure,imshow(a)
%converting an image to grayscale
I=rgb2gray(a);
%size of the image
[m,n]=size(I);
Totalcount=m*n;
%variables using to find the probability
cnt=1;
sigma=0;
%computing the cumulative probability.
for i=0:255
    k=I==i;
    count(cnt)=sum(k(:))
    %pro array is having the probabilities
    pro(cnt)=count(cnt)/Totalcount;
    sigma=sigma+pro(cnt);
    cumpro(cnt)=sigma;
    cnt=cnt+1;
end;
%Symbols for an image
symbols = [0:255];
%Huffman code Dictionary
dict = huffmandict(symbols,pro);
%function which converts array to vector
vec_size = 1;
for p = 1:m
    for q = 1:n
        newvec(vec_size) = I(p,q);
        vec_size = vec_size+1;
    end
end
%Huffman Encodig
hcode = huffmanenco(newvec,dict);
%Huffman Decoding
dhsig1 = huffmandeco(hcode,dict);
%convertign dhsig1 double to dhsig uint8
dhsig = uint8(dhsig1);
%vector to array conversion
dec_row=sqrt(length(dhsig));
dec_col=dec_row;
%variables using to convert vector 2 array
arr_row = 1;
arr_col = 1;
vec_si = 1;
for x = 1:m
    for y = 1:n
        back(x,y)=dhsig(vec_si);
        arr_col = arr_col+1;
        vec_si = vec_si + 1;
    end
    arr_row = arr_row+1;
end
%converting image from grayscale to rgb
[deco, map] = gray2ind(back,256);
RGB = ind2rgb(deco,map);
imwrite(RGB,'decoded.JPG');

Answer 2

[Abdel Rahman Bekawi]

well done, your code is great!

however, the thing that makes this logical error is that your map channels are equal to each other, hence you will get grayed scale image as well, so to overcome this issue, I recommend you work with indexed images.

% try the following 
[indexed, colormap] = rgb2ind(rgbimage, number_of_colors);
% do all the processing on this indexed format and then at the end 
RGB = ind2rgb(back, colormap);
% by this you should get a compressed colored image 

I hope this helps!

[Walter Roberson]

Note that if you are calculating compression efficiency or are expecting to be able to save the compressed image and restore it at a later time, then you should also be storing the rgb table and the Huffman dictionary

[hamza munawar]

have anyone done the huffman encoding on a 2D image in matlab???

[Walter Roberson]

Yes, people have done that. Just reshape the image into a vector, and Huffman encode that. If you save or transmit the encoded image, remember to store the dictionary and the image size so that you can reconstruct afterwards.

Answer 3

[Idin Motedayen-Aval]

This page is still getting a fair number of views, so I wanted to summarize the discussion.

• As it has been pointed out, the problem in the original code is the line "I = rgb2gray(a);". This line converts the image to grayscale, so everything is grayscale from that point on. The color information is lost and cannot be recovered.
• The right approach is to use indexed images as pointed out by @abdel rahman.
• If you want to store/transmit the compressed image, as @Walter Roberson pointed out a few times, "you should also be storing the rgb table and the Huffman dictionary."
• The OP's code is a mere academic exercise in Huffman coding an image, and then decoding it to recover the original image (since Huffman coding is lossless, the recovered image will be indentical to the original).
• I cleaned up the code a bit to show how it might run in MATLAB (it can be cleaned up much more; I preserved much of the original code for traceability)

%clearing all variableas and screen
clear all;
close all;
clc;
number_of_colors = 256;

%Reading image
a=imread('peppers.png');
figure(1),imshow(a)

%converting an image to grayscale
% I=rgb2gray(a);
% Use indexed image instead of grayscale
[I, myCmap] = rgb2ind(a, number_of_colors);

%size of the image
[m,n]=size(I);
Totalcount=m*n;

%variables using to find the probability
cnt=1;
% sigma=0;

%computing the cumulative probability.
pro = zeros(256,1);
for i=0:255
    k=(I==i);
    count=sum(k(:));
    %pro array is having the probabilities
    pro(cnt)=count/Totalcount;
    % sigma=sigma+pro(cnt); <-- Not needed here
    % cumpro(cnt)=sigma; <-- Not needed here
    cnt=cnt+1;
end
% Probablities can also be found using histcounts
pro1 = histcounts(I,0:256,'Normalization','probability');

cumpro = cumsum(pro); % if the cumulative sum is needed
sigma = sum(pro); % if the sum is needed; should always be 1.0

%Symbols for an image
symbols = 0:255;

%Huffman code Dictionary
dict = huffmandict(symbols,pro);

%function which converts array to vector
newvec = reshape(I,[numel(I),1]);

%Huffman Encodig
hcode = huffmanenco(newvec,dict);

%Huffman Decoding
dhsig1 = huffmandeco(hcode,dict);

%convertign dhsig1 double to dhsig uint8
dhsig = uint8(dhsig1);

%vector to array conversion
back = reshape(dhsig,[m n]);

%converting image from grayscale to rgb
% [deco, map] = gray2ind(back,256);
% RGB = ind2rgb(deco,map);
RGB = ind2rgb(back,myCmap);
imwrite(RGB,'decoded.JPG');
figure(2),imshow(RGB)
%end of the huffman coding

[Walter Roberson]

[I, myCmap] = rgb2ind(a, number_of_colors);

Note that with this approach of converting to indexed image, the restored image will not generally be identical to the original image. To have the restored image be identical to the original image, reshape the image from being m x n x 3 to being m x (n*3) and then encode that effectively-grayscale result.