MATLAB CODES - image segmentation based on Thresholding

clc;
close all;
clear all;

a = imread('grayflower256.jpg');
a = rgb2gray(a);
subplot(3,3,1);
imshow(a); title('Original Image');

% Simple thresholding at 0.3
% This is equivalent to saying threshoding at pixel value
% 0.3 x 255 = 76.5 , approximatesly 77
% imlabel() is a Matlab command that can threshold any matrix


level = 0.3;
%% Display the threshold image
subplot(3,3,2);
segimage1 = im2bw(a,level);
imshow(segimage1); title('Simple Thresholding at 0.3');


% Simple thresholding at 0.6
% So threshold value is 0.6 x 255 = 153
% Single Thresholding can be done like this also


%% Display the threshold image
subplot(3,3,3);
imshow(a > 153); title('Simple Thresholding at 0.6');


% Multiple thresholding Algorithm
% Let us assume that the output should be zero if pixel value is
% if <= 0.1 x 255 = 25.5 = 26, pixel output 204 if pixel value is <= 0.9 x
% 255 = 230 and 0 if pixel value is above 230.

%% Create a temporary matrix g

tmp = a;

[m n]= find(a<26);
for j = 1: length(m)
    tmp(m(j),n(j))=0;
end

[m n]= find(a>26 & a <= 230);
for j = 1: length(m)
    tmp(m(j),n(j))=0.8;
end

[m n]= find(a>230);
for j = 1: length(m)
    tmp(m(j),n(j))=0;
end
subplot(3,3,4);
segimage2 = im2bw(tmp,0);
imshow(segimage2); title('Multiple threshoding(Between 26-230)');

%% Find the threshold Value using Otsu Algorithm

level = graythresh(a);

%% Display the threshold image

subplot(3,3,5);
segimage = im2bw(a,level);
imshow(segimage); title('Otsu - Optimal Segmented Image');

%% Display Blured Image

b = imread('bluredtxt.jpg');
subplot(3,3,6);
imshow(b); title('Badly illuminated Image');

level = graythresh(b);
subplot(3,3,7);
segimage = im2bw(b,level);
imshow(segimage); title('Otsu - Segmentation for bad illuminated Image');


b = imread('bluredtxt.jpg');
b = rgb2gray(b);

%Create an average Image

avgfilt = ones(13,13);
adaptfiltmask = avgfilt/sum(avgfilt);
im = imfilter(b,adaptfiltmask,'replicate');

%Create an median image
im1 = medfilt2(b,[20 20]); 

%Adaptive threshold algorithm use
% threshold = mean + constant (Here 18)

thresh = im+18;
adaptthreshimg = b - thresh;
subplot(3,3,8);
imshow(adaptthreshimg > 0);


%Adaptive threshold algorithm used threshold = mean + constant (Here 2)
thresh1 = im1 + 2;
adaptthreshimg = b - thresh1;
subplot(3,3,9);
imshow(adaptthreshimg > 0);