how can make a dataset of images?

49 ビュー (過去 30 日間)
jack nn
jack nn 2015 年 6 月 14 日
コメント済み: Walter Roberson 2017 年 2 月 18 日
I provided a code that gets an initial image as input and gives me 4 images as result. I should run this code as loop that user enters an initial images. in this loop every time I should save all the results because I want to make a dataset of these images and then do feature extraction. how can I do this? Can I use structure for this? I think that we can imagine every image as a matrix but I don't how we can do this .I think I should have a dataset of images like this:
but I dont know how can make this.do you think this idea is rational?or I should change my approach.
I appreciate your helps and supports in advanced.

採用された回答

Walter Roberson
Walter Roberson 2015 年 6 月 14 日
A struct is fine.
datastruct = struct('IMG1', [], 'IMG2', [], 'IMG3', [], 'IMG4', []);
while true
[filename, pathstr] = uigetfile('Select an image');
if ~ischar(filename) %user canceled
break;
end
fullname = fullfile(pathstr, filename); %it might not be in current directory
... process file named by fullname
datastruct(end+1).IMG1 = image1;
datastruct(end).IMG2 = image2;
datastruct(end).IMG3 = image3;
datastruct(end).IMG4 = image4;
end
  1 件のコメント
jack nn
jack nn 2015 年 6 月 15 日
thanks dear Walter Roberson .

サインインしてコメントする。

その他の回答 (2 件)

Image Analyst
Image Analyst 2015 年 6 月 15 日
Why can't you do feature extraction immediately after you create them, and not worry about saving these intermediate images? Why do you have to save them all? If, for some reason you do, then I'd just save them in .mat files until you need them at some later time.
  7 件のコメント
Image Analyst
Image Analyst 2015 年 12 月 14 日
Alternatively, see this to get up to speed quickly with a nice GUI: http://www.mathworks.com/matlabcentral/fileexchange/24224-magic-matlab-generic-imaging-component

サインインしてコメントする。


mukul jain
mukul jain 2017 年 2 月 18 日
algorithm for identifying near surface/sub surface targets using hyperspectral image through Matlab. please guide me
  1 件のコメント
Walter Roberson
Walter Roberson 2017 年 2 月 18 日
Please create a new Question for that topic; it has nothing to do with this Question.

サインインしてコメントする。

タグ

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by