How to use adapthisteq on an AVI video

I am trying to increase the contrast of the frames within my AVI video file. I understand that adapthisteq can be used to increase contrast of single images so i want to adapt this code to cycle through each frame of my video and apply the contrast enhacement to each frame. I need the output to be an AVI video of equal resolution to the input AVI. Below is the code I have adapted from adaphisteq but I am running into the following errors. Any help would be appreciated. Thanks.

my code:

inputFullFileName = fullfile('file_x.avi');

% opening the videoreader for reading an input video file

inputVideoReaderObject = VideoReader(inputFullName)

% creating a videowriter object to write the video out to a new, different file

outputFullFileName = fullfile(pwd, 'file_x.avi');

outputVideoWriterObject = VideoWriter(outputFullFileName);

open(outputVideoWriterObject);

%running the adapthiseq code from: https://uk.mathworks.com/help/images/ref/adapthisteq.html

[X, MAP] = imread('file_x.avi');

RGB = ind2rgb(X, MAP);

LAB = rgb2lab(RGB);

L = LAB(:,:,1)/100;

L = adapthisteq(L, 'NumTiles', [8 8], 'ClipLimit', 0.005);

LAB(:,:,1) = L*100;

J = lab2rgb(LAB);

figure

imshowpair(RGB,J,'montage')

% loop through the movie, writing all frames out

for frame = 1 : numberOfFrames

% extracting the frame from the movie structure

thisInputFrame = read(inputVideoReaderObject, frame);

end

errors:

Error using imread>get_format_info

Unable to determine file format.

error in imread

fmt_s = get_format_info(fullname);

error in contrast_enhancer

[X, MAP] = imread('file_x.avi')

Answer 1

[Image Analyst]

Yes good idea but you have a number of errors in there. Mainly that your histogram equalization is not inside your loop over all frames. See my attached demo.

% Demo to extract frames from a input video, and build a new movie where each frame is histogram equalized.
% Illustrates the use of the VideoReader and VideoWriter classes.
% A Mathworks demo (different than mine) is located here http://www.mathworks.com/help/matlab/examples/convert-between-image-sequences-and-video.html

%-------------------------------------------------------------------------------------------------------------------------------------------
% Initialization steps
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
imtool close all;  % Close all imtool figures.
clear;  % Erase all existing variables.
workspace;  % Make sure the workspace panel is showing.
fontSize = 15;

%-------------------------------------------------------------------------------------------------------------------------------------------
% Open the rhino.avi demo movie that ships with MATLAB.
% First get the folder that it lives in.
folder = fileparts(which('rhinos.avi')); % Determine where demo folder is (works with all versions).
% Pick one of the two demo movies shipped with the Image Processing Toolbox.
% Comment out the other one.
inputFullFileName = fullfile(folder, 'rhinos.avi');
% movieFullFileName = fullfile(folder, 'traffic.avi');
% Check to see that it exists.
if ~exist(inputFullFileName, 'file')
    strErrorMessage = sprintf('File not found:\n%s\nYou can choose a new one, or cancel', inputFullFileName);
    response = questdlg(strErrorMessage, 'File not found', 'OK - choose a new movie.', 'Cancel', 'OK - choose a new movie.');
    if strcmpi(response, 'OK - choose a new movie.')
        [baseFileName, folderName, FilterIndex] = uigetfile('*.avi');
        if ~isequal(baseFileName, 0)
            inputFullFileName = fullfile(folderName, baseFileName);
        else
            return;
        end
    else
        return;
    end
end

%-------------------------------------------------------------------------------------------------------------------------------------------
% Open up the VideoReader for reading an input video file.
inputVideoReaderObject = VideoReader(inputFullFileName)
% Determine how many frames there are.
numberOfFrames = inputVideoReaderObject.NumberOfFrames;
inputVideoRows = inputVideoReaderObject.Height
inputVideoColumns = inputVideoReaderObject.Width

%-------------------------------------------------------------------------------------------------------------------------------------------
% Create an output VideoWriter object to write the video out to a new, different file.
outputFullFileName = fullfile(pwd, 'NewRhinos.avi');
outputVideoWriterObject = VideoWriter(outputFullFileName);
open(outputVideoWriterObject);
% Specify the output video size.
shrinkFactor = 4; % Shrink by a factor of 4 in both directions.
outputVideoRows = round(inputVideoRows / shrinkFactor)
outputVideoColumns = round(inputVideoColumns / shrinkFactor)
numberOfFramesWritten = 0;
% Prepare a figure to show the images.
hFig = figure('Name', 'Demo by Image Analyst', 'NumberTitle', 'off');

%-------------------------------------------------------------------------------------------------------------------------------------------
% Loop through the movie, writing all frames out.
for frame = 1 : numberOfFrames
    % Extract the frame from the movie structure.
    thisInputFrame = read(inputVideoReaderObject, frame);
    
    % Histogram equalize the image.
    % First convert the image to LAB color space because to make it look natural
    % we want to only filter the overall brightness, not each individual color channel.
    LAB = rgb2lab(thisInputFrame);
    L = LAB(:, :, 1); % Get the brightness or luminance channel.
    % L is floating point in the range 0-100.  If we are to use adapthisteq,
    % it must be in the range 0-1.
    L = L / 100;
    % Do the histogram equalization on this frame.
    L = adapthisteq(L, 'NumTiles', [8 8], 'ClipLimit', 0.005);
    % Undo the scaling to get it back into its original range.
    L = L * 100;
    % Put the filtered result back into the LAB array.
    LAB(:, :, 1) = L;
    % Turn it from LAB color space back into RGB colorspace.
    outputFrame = 255 * lab2rgb(LAB);
    % Make it uint8 so we can put side by side with the original frame.
    outputFrame = uint8(outputFrame);
    imshowpair(thisInputFrame, outputFrame,'montage');
    progressIndication = sprintf('Processed frame %4d of %d.', frame, numberOfFrames);
    title(progressIndication, 'FontSize', fontSize);
    drawnow;
    
    % Write this new, resized frame out to the new video file.
    writeVideo(outputVideoWriterObject, outputFrame);
    
    % Update user with the progress.  Display in the command window.
    progressIndication = sprintf('Processed frame %4d of %d.', frame, numberOfFrames);
    disp(progressIndication);
    % Increment frame count (should eventually = numberOfFrames
    % unless an error happens).
    numberOfFramesWritten = numberOfFramesWritten + 1;	
end
% Close the output video object.  You don't need to close the input video reader.
close(outputVideoWriterObject);

%================================================================================
% NOW WE'RE DONE!
%================================================================================
% Alert user that we're done.
finishedMessage = sprintf('Done!  It processed %d frames of\n"%s" and created output video\n%s.\nClick OK to see the output video.', numberOfFramesWritten, inputFullFileName, outputFullFileName);
fprintf('%s\n', finishedMessage); % Write to command window.
uiwait(msgbox(finishedMessage)); % Also pop up a message box.
% OPTIONAL: Play the movie in the default Windows program that plays videos.
winopen(outputFullFileName);

[David Gill]

thanks a lot! I have just run your code, it executes what i wanted to do. However, the processing time is quite long. Do you have any ideas how I could improve this? its, not the end of the world but anything to optimise the code further is always useful. thanks again.

[Image Analyst]

You could skip the displaying of each frame.

If it solved your question, could you click the "Accept this answer" link? Thanks in advance. 🙂

[David Gill]

Hi, i commented out line 83-86 so the frames werent displayed live. However, this has not changed the processing time; it is working approx 1 frame per second. Any other suggestions? thanks for all the help, I have marked you answer as you requested :)

[Image Analyst]

For me the demo took 3.7 seconds without the display lines and 10.7 (three times longer) with the display lines in there. I can't think of any other way to speed it up.

[David Gill]

did you remove the 'progress indication' for the command window as well? or did you just remove the progress indication for displaying the live frames (like I did)?

[Image Analyst]

I don't remember but it's not needed. If you don't want it, comment it out. Printing stuff to the command window will slow down the code. Sounds like you have a super long video though and it takes a very long time. So in that case I'd just use rem or mod to print out every 1000th frame number because with a video that long you don't want no indication of how far along it is. You'd just think it got hung if you saw no progress for several minutes.

[David Gill]

assuming I would change the following line of code to acheive this? I changed 1 to 10 to test if it displayed every ten frames instead but still displaying count with an increment of 1.

proressIndication = sprintf('processed frame %4 of %4.', frame, numberOfFrames);

disp(progressIndication);

%increment frame count (should eventually = numberOfFrames unless error happens)

numberOfFramesWritten = numberOfFramesWritten + 1;

[Image Analyst]

I don't see anything there that executes those lines only every 10 frames. Where did you use rem or mod like I suggested? You need to do

numberOfFrames = 100;
for frame = 1 : numberOfFrames
    if rem(frame, 10) == 0 % Only print every 10th frame
        progressIndication = sprintf('Processed frame #%4d of %4d.', frame, numberOfFrames);
        disp(progressIndication);
    end
end

Processed frame #  10 of  100.
Processed frame #  20 of  100.
Processed frame #  30 of  100.
Processed frame #  40 of  100.
Processed frame #  50 of  100.
Processed frame #  60 of  100.
Processed frame #  70 of  100.
Processed frame #  80 of  100.
Processed frame #  90 of  100.
Processed frame # 100 of  100.