thresholding_estimation.m

clc
clear all
close all

I = imread('/Users/ehzawad/Desktop/footBall_orig.jpg');
% I = imread('peppers.png');

Im = rgb2gray(I);

% disp(Im);

% I = imread('bag.png');



%
% figure, subplot(2,2,1), imshow(I), title('Original Image');
% [I_prw1,t1] = edge(I,'Canny'); % % sobel prewitt Roberts log zerocross Canny
% subplot(2,2,2), imshow(I_prw1);
% I_noise = imnoise(I,'gaussian');
% % [I_prw2,t2] = edge(I_noise,'Canny'); % sobel prewitt Roberts log zerocross Canny
% [I_prw2,t2] = edge(I_noise,'Canny', [], 2.5); % sobel prewitt Roberts log zerocross Canny
% subplot(2,2,3), imshow(I_noise);
% subplot(2,2,4), imshow(I_prw2);


th = .50;

% I_bw = im2bw(I,.5);

% I_bw = imbinarize(I, 0.5);

% iteration = 1000;

flattenImage = Im(:);

% disp(flattenImage);

% iteration = length(I_bw);

for i = 1 : iteration
    %first segment or piece
    disp(th);
    G1 = flattenImage > th;
    disp(G1);
    G2 = flattenImage <= th;
    disp(G2);
    median = (mean(G1) + mean(G2))/ 2;
    % disp(median);

    if (abs(median - th) <= 0.001)
        break
    end

    th = median;
    % disp(th);

end

newEstimated = im2bw(I, th);

 imshow(newEstimated);