Error using vertcat Dimensions of arrays being concatenated are not consistent.

My problem is I am trying to compress and decompress an image file using Huffman Coding,

here is the code where I get the error "Error using vertcat Dimensions of arrays being concatenated are not consistent." I undelined the part where the code is not running

% Compress function

function compressed_data = compress_image(image)

% Convert the image to a 1D array

image = image(:);

% Calculate the frequency of each pixel value

freq = histcounts(image, unique(image));

% Create a cell array to store the pixel values and their corresponding frequencies

pixel_freq = num2cell(freq);

pixel_val = num2cell(unique(image));

pixel_data = [pixel_val; pixel_freq];

% Sort the pixel data based on frequency

[~, sort_index] = sort(freq, 'descend');

sorted_pixel_data = pixel_data(:, sort_index);

% Generate the Huffman code dictionary

huffman_dict = generate_huffman_dict(sorted_pixel_data);

% Encode the image using Huffman coding

encoded_data = encode_image(image, huffman_dict);

% Pack the compressed data as a struct

compressed_data = struct('image_size', size(image), 'huffman_dict', huffman_dict, 'encoded_data', encoded_data);

end

Answer 1

[Image Analyst]

pixel_val and pixel_freq have different number of columns -- check it for yourself -- so you cannot stitch them together vertically, one atop the other.

Also, "image" is a built-in function so you must not destroy it by using it as a variable name.

Answer 2

[Walter Roberson]

histcounts() returns row vectors whereas unique() of a non-row returns a column vector. (Unique of a row vector returns a row vector.)

You used

 image = image(:);

so image is not a row vector, so unique() is going to return a column vector.

With histcounts returning a row vector and unique returning a column vector, vertcat of the two cannot work.

[Sandra]

thank you I have a better understanding now, is there a function instead of unique() that would solve this issue?

[Image Analyst]

There is no reason for you to do that in the first place. It is simply not needed.

[Sandra]

Can you please explain why is not needed? I need to convert the image into 1D array and I also need the values of the pixels and their frequecies in order to calculate the probability and sort based on Huffman algorithm. Would appreciate if you could explain your answer :)