Batch process AVI files into PNG/JPG

Hi all,

I am working with a large dataset of .avi files which I need to split into .png for further analysis. From previous questions on here, I've derived some code which is doing this quite nicely for one avi file:

clc;
clear;

%Set current directory to folder containing this m file
cd('C:\Users\...');

%Open file to be converted
movieFullFileName = fullfile(cd, '26032019.AVI'); 

%Read in avi file
videoObject = VideoReader(movieFullFileName) 

%Get some info from avi file
[folder, baseFileName, extentions] = fileparts(movieFullFileName); 

%Create output folder:
folder=pwd;
outputFolder = sprintf('%s/Frames', folder); 

%Determine how many frames there are
numberOfFrames = videoObject.NumberOfFrames;
vidHeight = videoObject.Height;
vidWidth = videoObject.Width;


for x=1 : numberOfFrames
frame = read(videoObject,x)
outputBaseFileName = sprintf('-%4.4d.png', x); %output filename, each frame numbered from 0001
outputFullFileName = fullfile(outputFolder, strcat(baseFileName, outputBaseFileName)); %output filename
imwrite(frame, outputFullFileName, 'png'); %write output file
end

At the moment, I'm only converting one video file and it takes just over a minute to process which seems quite long. Once I am finished collecting my data I'll have over 1500 videos so I am wondering if, at this early stage, there is anything I can do to refine or speed up my code?

Secondly, I would like to introduce another for loop which lets me apply this code to a series of avi files within a folder. Can anyone point me in the right direction with this?

Thanks!!

[Louise Wilson]

Hi everyone, I have managed to modify the code to include a for loop:

%%08 July 2019, Louise Wilson
%Code to split .avi file into many .jpg files.
%Same code but attempting for loop to analyse multiple files in a folder...

clc;
clear;

%Set current directory to folder containing this m file
paths=cd('C:\Users\lwil634\Documents\Cameras\Test');


for i=1:length(paths)
    d=dir(fullfile(paths, '*.avi'));

    for j=1:length(d)
    vidfile=d(j).name
       
    videoObject=VideoReader(vidfile)
    
    [folder, baseFileName, extentions] = fileparts(vidfile); 

    folder=pwd;
    outputFolder = sprintf('%s/Frames', folder); 

    numberOfFrames = videoObject.NumberOfFrames;
    vidHeight = videoObject.Height;
    vidWidth = videoObject.Width;


    for x=1 : numberOfFrames
    frame = read(videoObject,x)
    outputBaseFileName = sprintf('-%4.4d.png', x); %output filename, each frame numbered from 0001
    outputFullFileName = fullfile(outputFolder, strcat(baseFileName, outputBaseFileName)); %output filename
    imwrite(frame, outputFullFileName, 'png'); %write output file
    end
    
    end
end

This code works-I get all of the frames for each AVI file in the folder I specify. However, I think it is erroneous? When I run the code, it runs forever, and doesn't stop until I ctrl+c. I only did this once I checked the output folder and saw that all of the files I expected were there. What can do to stop the code running on and on even though I believe it to be finished...?

Ultimately I am trying to write something so that I can put multiple paths in the paths section, so I could use the script to run through many folders of avi files at one execution of code.

Thanks!!

[KALYAN ACHARJYA]

it runs forever, and doesn't stop until I ctrl+c.

Please check on

• length(paths)
• length(d)
• numberOfFrames

Do confirm?

[Edric Ellis]

I think what's going wrong here is the outer-most loop. This code:

paths = cd('...');

returns in paths a char-vector of the old working directory, which will be something like 'c:\my\old\working\directory'. Because paths is a vector of type char, the outermost loop is simply repeating the processing multiple times doing exactly the same thing each time - once per character in paths. This is not what you want! To fix, simply omit the outermost loop.

I'm not familiar with video processing - but I am familiar with parallel computing - so one thought might be (if you have Parallel Computing Toolbox available) to convert the loop over the video files to be a parfor loop. Therefore, your code might look like this:

videoDir = 'C:\Users\lwil634\Documents\Cameras\Test';
videoFiles = dir(fullfile(videoDir, '*.avi'));
parfor fIdx = 1:numel(videoFiles)
    thisVideoFile = fullfile(videoDir, videoFiles(fIdx).name);
    % process thisVideoFile...
end

[Louise Wilson]

Thank you Edric! I did what you said and removed the outer loop, and now it works without problems, and is much faster!!

However, now I am only looping through multiple files within one folder. If possible, I was hoping to loop through many folders/paths? So, for example, I would list a series of C:/ destinations within paths. And then the loop would loop through these paths... does it make sense?

[Louise Wilson]

Hi Edric, I'm also curious about the code as it plays back the video to get each frame, am I correct? Is it possible to do similar but without this... so to speed up the code? I am looking at readframe but not understanding it:https://au.mathworks.com/help/matlab/ref/videoreader.readframe.html

[Edric Ellis]

To loop over multiple folders, the simplest way is to use an outer loop like this:

videoFolders = {'c:\path\to\folder\one', 'c:\path\to\folder\two'};
for folderIdx = 1:numel(videoFolders)
    thisFolder = videoFolders{folderIdx};
    theseFiles = dir(fullfile(thisFolder, '*.avi'));
    for fileIdx = 1:numel(theseFiles)
        ... % process each file in the folder
    end
end

Another option if there's a common root folder, and you're using a recent version of MATLAB, is to use the '**' syntax in the dir command, like this:

commonVideoFolder = 'c:\root\path\to\videos';
allFiles = dir(fullfile(commonVideoFolder, '**', '*.avi'));
for fileIdx = 1:numel(allFiles) % could be 'parfor'
    thisEntry = allFiles(fileIdx);
    thisFileName = fullfile(thisEntry.folder, thisEntry.name);
    % ... process the file.
end

Answer 1

[Image Analyst]

See the FAQ: https://matlab.fandom.com/wiki/FAQ#How_can_I_process_a_sequence_of_files.3F

Put your existing code inside the for loop of the FAQ.

[Louise Wilson]

Hi! Thank you! I have my code working for lots of *.avi files... but I want to run through many folders of avi files. Should I adapt the code in the link you mention to include a series of folders?

% Specify the folder where the files live.
myFolder = 'C:\Users\yourUserName\Documents\My Pictures'; %other folders here???
% Check to make sure that folder actually exists.  Warn user if it doesn't.
if ~isdir(myFolder)
  errorMessage = sprintf('Error: The following folder does not exist:\n%s', myFolder);
  uiwait(warndlg(errorMessage));
  return;
end
% Get a list of all files in the folder with the desired file name pattern.
filePattern = fullfile(myFolder, '*.jpg'); % Change to whatever pattern you need.
theFiles = dir(filePattern);
for k = 1 : length(theFiles)
  baseFileName = theFiles(k).name;
  fullFileName = fullfile(myFolder, baseFileName);
  fprintf(1, 'Now reading %s\n', fullFileName);
  % Now do whatever you want with this file name,
  % such as reading it in as an image array with imread()
  imageArray = imread(fullFileName);
  imshow(imageArray);  % Display image.
  drawnow; % Force display to update immediately.
end

My code right now:

%%08 July 2019, Louise Wilson
%Code to split several .avi files within a folder into many .jpg files.

%Adapting code to use vision.VideoFileReader might be faster?

clc;
clear;

%Set current directory to folder containing this .m file and raw .avi files:
paths=cd('C:\Users\lwil634\...'); 

mkdir Frames %create Frames output folder

d=dir(fullfile(paths, '*.AVI')); %filename of each file to be processed

fileCount=length(d); %number of files to be processed
    

for j=1:fileCount
    vidfile=d(j).name; %get name of jth video file
       
    videoObject=VideoReader(vidfile); %read in avi file
    
    [folder, baseFileName, extentions] = fileparts(vidfile); %get info from avi file

    folder=pwd; %print working directory
    outputFolder = sprintf('%s/Frames', folder); %put outfiles files in 'Frames' folder

    %Determine how many frames there are
    numberOfFrames = videoObject.NumberOfFrames;
    vidHeight = videoObject.Height;
    vidWidth = videoObject.Width;


    for x=1 : numberOfFrames
    frame = read(videoObject,x);
    outputBaseFileName = sprintf('-%4.4d.png', x); %output filename, each frame numbered from 0001
    outputFullFileName = fullfile(outputFolder, strcat(baseFileName, outputBaseFileName)); %full output filename with .avi file number
    imwrite(frame, outputFullFileName, 'png'); %write output png file
    end
    
end