Difficulty in applying huffman encoding for image compression

I got the given code to apply Huffman Encoding for image compression. But i am getting the error-

Error using huffmandict (line 71)
The symbol input must be a vector
Error in new (line 5)
[dict,avglen]=huffmandict(symbols,p)

The code is:

A=imread('xyz.jpg');
[symbols,p]=hist(A,double(unique(A)))
p=p/sum(p)
[dict,avglen]=huffmandict(symbols,p)
comp=huffmanenco(A,dict)

Help me as i am new to all these concepts.

Answer 1

[KALYAN ACHARJYA]

When I have tested your code with the test image, both symbols and p have different sizes (see in the attached image) Read the documentation about huffmandict Clearly mentioned that-

The length of p must equal the length of symbols.

-------------------------------------------------------------------------------------------------------------------------------------------- Updated Answer

I have suspected another issue if you consider any 1 D array, the following code is working (Note that an image is 2D)

%A=double(imread('test.jpg'));
p=[.5 .125 .125 .125 .0625 .0625];
A=randsrc(100,1,[1:6; p]);
%[symbols,p]=hist(A,double(unique(A)));
[p,symbols]=hist(A,double(unique(A))); 
p=p/sum(p);
[dict,avglen]=huffmandict(symbols,p);
comp=huffmanenco(A,dict);

[Nidhi Kumari]

Now i have converted the image to vetor ,but new error is coming -

Source symbols repeat

Also the new 'p' is 256x1 double and 'symbols' is 1x256 double. What should i do now?

[KALYAN ACHARJYA]

I have updated the answer, please check.

Have you read here?

[Nidhi Kumari]

I am using R2014a ,so histcounts() is not present. Can you suggest any other alternative?

Answer 2

[Walter Roberson]

 [symbols,p] = hist(A(:), double(unique(A)));

[Nidhi Kumari]

Now new error is coming-

Source symbols repeat

[Walter Roberson]

[symbols, idx] = unique(A(:));
counts = accumarray(idx, 1);
p = counts ./ sum(counts);
[dict,avglen]=huffmandict(symbols,p);
comp=huffmanenco(A,dict);

[Nidhi Kumari]

Error-

The symbol and probability vector must have the same length

[Walter Roberson]

[symbols, ~, idx] = unique(A(:));
counts = accumarray(idx, 1);
p = counts ./ sum(counts);
[dict,avglen]=huffmandict(symbols,p);
comp=huffmanenco(A(:),dict);

[Nidhi Kumari]

The output image is a vertical thin line along with the message-

Warning: Image is too big to fit on screen; displaying at 0%

[Walter Roberson]

The output is not an image: it is a double vector containing the values 0 and 1.