I want to read multiple files from a folder, the only problem in my code down is that
instead of having the full image path:
'checkboard/Gaussian_Noised/image_01.jpg'
I have the path without the zero in the last two digits
EDIT: There is space instead of zero on the first 9 Image paths.
'checkboard/Gaussian_Noised/image_ 1.jpg'
how can I fix this please
%-------------------------------------------------------
clc
clear all
close all
%------------------------------------------------------
path = 'checkboard/Gaussian_Noised/';
% Define images to process
imageFileNames = cell(1,36);
for n = 1:36
imageFileNames{n} = strcat(path,sprintf('image_%2d.jpg',n))
end
%------------------------------------------------------
In this case, it is better to use the float-format-specifier to pad zeros:
num2str(1,'%02.f')
'01'
num2str(1,'%02.d')
'1'
num2str(1,'%2.d')
'1'
num2str(1,'%d')
'1'
So in your case:
sprintf('image_%02.f.jpg',n)
general advice
As a general advice, you may want just to check the folder for all files of a certain pattern
path2dir = pwd; % path to working directory
pattern = 'image_*.jpg'
Lst = dir(fullfile(path2dir ,pattern));
for i = 1:length(Lst)
path2file = fullfile(Lst(i).folder,Lst(i).name);
% load the file and do something with it
end
Related
I have a set of files in a folder that I want to convert to a different type using MATLAB, ie. for each file in this folder do "x".
I have seen answers that suggest the use of the "dir" function to create a structure that contains all the files as elements (Loop through files in a folder in matlab). However, this function requires me to specify the file type (.csv, .txt, etc). The files I want to process are from a very old microscope and have the file suffixes .000, .001, .002, etc so I'm unable to use the solutions suggested elsewhere.
Does anyone know how I can get around this problem? This is my first time using MATLAB so any help would be greatly appreciated. All the best!
As suggested by beaker and Ander Biguri, you can do this by getting all files and then sorting.
Here is the related code with explanations inline:
baseDir = 'C:\mydir';
files = dir(baseDir);
files( [ files.isdir] ) = []; % This removes directories, especially '.' and '..'.
fileNames = cell(size(files)); % In this cell, we store the file names.
fileExt = cell(size(files)); % In this cell, we store the file extensions.
for k = 1 : numel(fileNames )
cf = files(k);
fileNames{k} = [ cf.folder filesep cf.name]; % Get file name
[~,~, fileExt{k}] = fileparts( cf.name ); % Get extension
end
[~,idx] = sort( fileExt ); % Get index to sort by extension
fileNames = fileNames( idx ); % Do the actual sorting
I have a folder origin_training with subfolders like FM, folder1, folder2, ... . I can get the list of image files in .png format into a cell called FM.
FM_dir = fullfile(origin_training, 'FM');
FM = struct2cell(dir(fullfile(FM_dir, '*.png')))';
My goal is to match the names in my folder with the images in my cd, and make a new folder FM with the images from cd. Image names in the two paths are identical. I can do that with:
% mother_folder = '...' my current directory
filenames = FM(:,1);
destination = fullfile(mother_folder, 'FM', FM(:,1));
cellfun(#copyfile,filenames, destination);
This is really slow. Takes minutes even for small amounts of images (~30).
My current directory has ~ 10000 images (the ones that correspond to FM, folder2, folder3, ...). What am I missing here?
An alternative would be to create a shell command and execute it. Assuming FM contains full paths for each file, or relative paths from the current directory, then:
FM_dir = fullfile(origin_training, 'FM');
destination = fullfile(mother_folder, 'FM');
curdir = cd(FM_dir);
FM = dir('*.png');
cmd = ['cp ', sprintf('%s ',FM.name), destination];
system(cmd);
cd(curdir);
If you're on Windows, replace 'cp' by 'copy'.
Note that here we're not creating a shell command per file to be copied (presumably what copyfile does), but a single shell command that copies all files at once. We're also not specifying the name for each copied file, we specify the names of the files to be copied and where to copy them to.
Note also that for this to work, mother_folder must be an absolute path.
I somehow put my code into a function and now it works at expected speed. I suspected that cd had to do with the low speed so I am only passing the full directories as character vectors.
The same approach works for my purpose or with a slight modification to just copy from A to B. As now, it works to match files in A to those on B and copy to B/folder_name
function [out] = my_copyfiles(origin_dir, folder_name, dest_dir, extension)
new_dir = fullfile(origin_dir, folder_name);
% Get a cell with filenames in the origin_dir/folder_name
% Mind transposing to have them in the rows
NEW = struct2cell(dir(fullfile(new_dir, extension)))';
dest_folder = fullfile(dest_dir, folder_name);
% disp(dest_folder)
if ~exist(dest_folder, 'dir')
mkdir(dest_folder)
end
%% CAUTION, This is the step
% Use names from NEW to generate fullnames in dest_dir that would match
filenames = fullfile(dest_dir, NEW(:,1));
destination = fullfile(dest_folder, NEW(:,1));
% if you wanted from origin, instead run
% filenames = fullfile(new_dir, NEW(:,1));
% make the copies
cellfun(#copyfile,filenames, destination);
%Save the call
out.originDir = new_dir;
out.copyFrom = dest_dir;
out.to = dest_folder;
out.filenames = filenames;
out.destination = destination;
end
I have a cell array of full file names and I want to find the lowest common directory where it makes sense to store accumulated data and what not.
Here is an example hierarchy of test data:
C:\Test\Run1\data1
C:\Test\Run1\data2
C:\Test\Run1\data3
C:\Test\Run2\data1
C:\Test\Run2\data2
.
.
.
In Matlab, the paths are stored in a cell array as follows (each run shares a row):
filePaths = {...
'C:\Test\Run1\data1','C:\Test\Run1\data2','C:\Test\Run1\data3'; ...
'C:\Test\Run2\data1','C:\Test\Run2\data2','C:\Test\Run2\data3'};
I want to write a routine that outputs the common path C:\Test\Run1 so that I can store relevant plots in a new directory there.
C:\Test\Run1\Accumulation_Plots
C:\Test\Run2\Accumulation_Plots
.
.
.
Previously, I was only concerned with two files in an x-by-2 cell, so the regiment below worked; however, strcmp lost it's appeal since I can't (AFAIK) index the whole cell at once.
d = 1;
while strcmp(filePaths{1}(1:d),filePaths{2}(1:d))
d = d + 1;
end
common_directory = filePaths{1}(1:d-1);
mkdir(common_directory,'Accumulation_Plots');
As suggested by #nekomatic, I'm posting my comment as an answer.
filePaths = {...
'C:\Test\Run1\data1','C:\Test\Run1\data2','C:\Test\Run1\data3'; ...
'C:\Test\Run2\data1','C:\Test\Run2\data2','C:\Test\Run2\data3'};
% Sort the file paths
temp = sort(filePaths(:));
% Take the first and the last one, and split by '\'
first = strsplit(temp{1}, '\');
last = strsplit(temp{end}, '\');
% Compare them up to the depth of the smallest. Find the 'first N matching values'
sizeMin = min(numel(first), numel(last));
N = find(~[cellfun(#strcmp, first(1:sizeMin), last(1:sizeMin)) 0], 1, 'first') - 1;
% Get the smallest common path
commonPath = strjoin(first(1:N), '\');
You just need to compare the first d characters of any path in the array - e.g. path 1 - with the first d characters of the other paths. The longest common base path can't be longer than path 1 and it can't be shorter than the shortest common base path between path 1 and any other path.
There must be several ways you could do that, but a concise one is using strfind to match the strings and cellfun with isempty to check which ones didn't match:
% filePaths should contain at least two paths
filePaths = {...
'C:\Test\Run1\data1','C:\Test\Run1\data2','C:\Test\Run1\data3'; ...
'C:\Test\Run2\data1','C:\Test\Run2\data2','C:\Test\Run2\data3'};
path1 = filePaths{1};
filePaths = filePaths(2:end);
% find longest common left-anchored substring
d = 1;
while ~any(cellfun(#isempty, strfind(filePaths, path1(1:d))))
d = d + 1;
end
% find common base path from substring
[common_directory, ~, ~] = fileparts(path1(1:d));
Your code leaves d containing the length of the longest common left-anchored substring between the paths, but that might be longer than the common base path; fileparts extracts the actual base path from that substring.
I have 200 JPEG images numbered from 1200 to 1399. How do I change their names from 1200.jpg-1400.jpg to 1.jpg-200.jpg, while keeping the original order of the image names?
This is a problem which can be tackled more efficiently in bash or other shell scripting languages. Probably, it can even be solved just via a single shell find.
Anyway, matlab can also do the job.
Consider this.
list = dir('./stack*.png'); % # assuming the file names are like stack1200.jpg
offset = -1200; # % numbering offset amount
for idx = 1:length(list) % # we go through every file name
name = list(idx).name;
number = sscanf(name,'stack%f.png',1); % # we extract the number
movefile(name,['stack' num2str(number + offset) '.png' ]); % # we rename the file.
end
I have find this solution which is quiet straightforward.
Pathfold= ('Path\to\images\');
dirData = dir('path\to\images\*.jpg');
fileNames = {dirData.name};
for i = 1:size(fileNames)
newName = sprintf('%d.jpg',i);
movefile(fileNames{i},[Pathfold , newName]) %# Rename the file
end
What is the best way to figure out the size of a file using MATLAB? The first thought that comes to mind is size(fread(fid)).
Please see the dir function as stated above.
Please note that the dir function works on files and not on directories only.
>> s = dir('c:\try.c')
s =
name: 'try.c'
date: '01-Feb-2008 10:45:43'
bytes: 20
isdir: 0
datenum: 7.3344e+005
You can use the DIR function to get directory information, which includes the sizes of the files in that directory. For example:
dirInfo = dir(dirName); %# Where dirName is the directory name where the
%# file is located
index = strcmp({dirInfo.name},fileName); %# Where fileName is the name of
%# the file.
fileSize = dirInfo(index).bytes; %# The size of the file, in bytes
Or, since you are looking for only one file, you can do what Elazar said and just pass an absolute or relative path to your file to DIR:
fileInfo = dir('I:\kpe\matlab\temp.m');
fileSize = fileInfo.bytes;
Use the fact that MatLab has access to Java Objects:
myFile = java.io.File('filename_here')
flen = length(myFile)
If you don't want to hardcode in your directory, you can use the built in pwd tool to find the current directory and then add your file name to it. See example below:
FileInfo = dir([pwd,'\tempfile.dat'])
FileSize = FileInfo.bytes
The question seems to indicate that fopen/fread/.. is used. In this case, why not seeking to the end of the file and reading the position?
Example:
function file_length = get_file_length(fid)
% extracts file length in bytes from a file opened by fopen
% fid is file handle returned from fopen
% store current seek
current_seek = ftell(fid);
% move to end
fseek(fid, 0, 1);
% read end position
file_length = ftell(fid);
% move to previous position
fseek(fid, current_seek, -1);
end
Matlab could have provided a shortcut..
More on ftell can be found here.
This code works for any file and directory (no need for absolute path) :
dirInfo=dir(pwd);
index = strcmp({dirInfo.name},[filename, '.ext']); % change the ext to proper extension
fileSize = dirInfo(index).bytes
Easy way to find size of file is:
enter these cammands
K=imfinfo('filename.formate');
size_of_file=K.FileSize
and get size of file.