Tag Archives: image processing

[Matlab] Classifying Image based on Its Label

Posted on by
Reply

Let’s back in MATLAB code. Now, I would like give an example for classifying images based on its label. Below, we have six images and in each has label, i.e 1.JPG ( label 1), 2.JPG (label 1), 3.JPG (label 2), 4.JPG (label 2), 5.JPG (label 1), and 6.JPG (label 2). We’re going to cluster images which has label 1.

Screen Shot 2014-06-27 at 3.19.41 pm

Here the code of function of classifying images. Please save it into Cluster.m.

[code language=”matlab”]
% fungsi ini digunkaan untuk menampilkan gambar2 yang berada pada kelas
% yang sama
% folder = nama folder yang berisi gambar-gambar
% label = label dari masing-masing gambar
function Cluster(folder,label,kelas)

a=find(label==kelas); % membaca indeks klasifikasi sesuai dengan kelas yang diinputkan
files = dir([folder,’*.JPG’]);
m=struct2cell(files);

figure
for k=1:size(a,1) % looping sesuai banyaknya anggota dalam kelas yang diinputkan
file = [folder m{1,a(k)}]; % file yang dikelompokkan sesuai dengan kelasnya
readfile =imread(file); % membaca file
subplot(2,2,k), imshow(readfile), title ([‘kelas’,num2str(kelas),’-‘,num2str(a(k))]);
end
[/code]

And you may run this code below as Main. We would like to show all images which is placed in label 1 and the output of this Main is shown in Figure 1

[code language=”matlab”]
clear
clc
close all

folder = ‘daun/’; % nama folder
label = [1;1;2;2;1;2]; %nama label

Cluster(folder,label,1);
[/code]

Screen Shot 2014-06-27 at 3.16.16 pm

Well,  this phase may help you to cluster the images well. This code presented for @Evy Kamilah 🙂 Good luck Evy 🙂

[Matlab] How to Retrieve Color from Segmented Image?

Posted on by
Reply

Back to code, yeah! Still in the simple code. Yep, simple code that is still quite make several people not clearly enough hove to solve it. At the moment, we will discuss about how to retrieve or return the color of segmented image. This topic is still related to image processing or computer vision, so if you are not in that specific interest area, perhaps it is not too useful. However if you are one of the people who has high curiosity, it is not close opportunity to give inquire if there is anything you want to know.

In this case, for instance  I have an image as shown below. Do you know what is that? (Yep, tomato :D). Segmentation is mostly happened in the first step in image processing. We have to recognize the object of an image to be processed in the next phase. Otsu thresholding was used in this case, and you can modified of Otsu thresholding as you wish in order to upgrade the quality of the segmentation. As we know, there are a lot of algorithms nowadays that compete to achieve the best segmentation, and it is always an open case. Well, you can follow this code right now to try.

tom

[code language=”matlab”]
% Copyright by Yuita Arum Sari
% This function is able to retrieve or return original color from the
% foreground of an image
% Input : file name of the image (full path, if applicable)

function retrieveColor(fileImage)

% read image
I = imread(fileImage);

%segmented image with Otsu thresholding
level = graythresh(I);
BW = im2bw(I,level);

% Retrieve color from segmented image
dim = size(BW);

% initialization of an image
retrieved = zeros (dim);
[a b]= size(BW);
retrieved = imresize(I,[a b]);

Bw = ~BW; % inverse black and white
R = retrieved(:,:,1).* uint8(Bw);
G = retrieved(:,:,2).* uint8(Bw);
B = retrieved(:,:,3).* uint8(Bw);

% retrieved color
retrieved(:,:,1) = R;
retrieved(:,:,2) = G;
retrieved(:,:,3) = B;

figure
subplot(1,3,1), imshow(I), title (‘Original Image’);
subplot(1,3,2), imshow(Bw), title (‘Segmented Image’);
subplot(1,3,3), imshow(retrieved), title (‘Retrieved Color’);
[/code]

Have you finished to code it? Alright, you can event copy paste this function to try. Then, you can run code by this code

[code language=”matlab”]
retrieveColor(‘tom.jpg’)
[/code]

Screen Shot 2014-05-29 at 12.44.03 pm

From picture above, you can see original image, segmented image, and retrieved color, respectively. If you want to deal with gaining teh RGB value, you can add by your own :

[code language=”matlab”]
function [R,G,B]= retrieveColor(fileImage)
[/code]

Well, is it easy? 😀 Do you want anything to share related this post? I am welcoming of your comment. Thank you!  🙂

[Matlab] Read Multiple Images in a Folder

Posted on by
Reply

Hello guys,  in this occasion I would like to share about how to read multiple images in a folder, and save the information into array. This code is conducted by Matlab, that really approachable for graduate students (subjective opinion 😀 ) . I’ll present this code for my lovely friend Ratih, and perhaps is also helpful for you all, because this case is often happened when we talk about classification, clustering, information retrieval in image processing or computer vision.

Well, for instances I have a folder named “capturedImage” as shown in a figure below. In a folder contains 24 images (let’s say 24 different colors below are 24 images hehe..).

Screen Shot 2014-02-21 at 5.42.55 pm

Let’s read the information inside image into code that is described in a brief code below. In every line has its own narrative and may help you to understand what is the code indicate. I’ll take a case about how to get mean among intensities in every image, so I used data_mean to define collected information of all images in a folder through an array. Check this code out.

[code language=”matlab”]
% Cara ini digunakan untuk membaca gambar-gambar dalam satu folder
% dan kemudian meletakkkan nilai fitur gambar tersebut ke dalam sebuah
% array
% =====================================================================

clc
clear

gambar = ‘capturedImage/’; % nama folder "campturedImage"
% mengindetifikasi isi folder dalam file yang berekstensi .JPG
files = dir([gambar,’*.JPG’]);

m=struct2cell(files); % menentukan informasi gambar yang ada dalam folder
data_mean=[]; % inisialisasi awal array

for k=1:size(m,2)
file = [gambar m{1,k}]; % identifikasi nama file beserta path-nya
dataMean = mean(mean((file))) % menghitung mean dari masing-masing gambar
data_mean=[data_mean;dataMean]; % memasukkan data mean ke dalam array
end
[/code]

Ok, that’s all 🙂 As usual the open question to share is free welcome. Thank you and happy coding 🙂