MATLAB Answers

Is there a way to overcome For loops for fast processing?

1 view (last 30 days)
I have written a following MATLAB code to detect skin color pixels from the RGB image. The output is as expected but being pixel by pixel algorithm, my program is taking too long, expecially for high quality images. Is there a way to take out the for loops for fast processing? Thank you.
[s1,s2,s3] = size(I);
skin_img1 = zeros(s1,s2,s3);
maxR = max(max(R));
minR = min(min(R));
maxG = max(max(G))
minG = min(min(G));
maxB = max(max(B));
minB = min(min(B));
for i = 1:s1
for j = 1:s2
if R(i,j)>90/255
if ( G(i,j)>40/255 && B(i,j)>20/255 && abs(R(i,j)-G(i,j))>15/255 && R(i,j)> G(i,j) && R(i,j) > B(i,j)...
&& (max([R(i,j),G(i,j),B(i,j)])- min([R(i,j),G(i,j),B(i,j)]) >15/255));
% (maxR - minR)>15/255 && (maxG - minG)>15/255 && (maxB - minB)>15/255 )
skin_img1(i,j,:) = I (i,j,:);
end
% else
% if( R(i,j)> G(i,j) && R(i,j) > B(i,j) && R(i,j) - G(i,j)> 5/255 && (abs(G(i,j)-B(i,j))>5/255))
% Skin_Image(i,j,:) = I (i,j,:);
% end
end
end
end

Accepted Answer

Muhammad Farhan  Mughal
Muhammad Farhan Mughal on 10 Sep 2019
Following MATLAB code significantly reduced the computatonal cost of the questioned code
R = I(:,:,1);
G = I(:,:,2);
B = I(:,:,3);
R = double(R);
G = double(G);
B = double(B);
R = R/255;
G = G/255;
B = B/255;
[s1,s2,s3] = size(I);
skin_img1 = zeros(s1,s2,s3);
index1 = zeros(size(R));
R1 = R(:) ; G1 = R(:); B1 = B(:);
mat_I = [R1,G1,B1]';
index = R >90/255 & G > 40/255 & B>20/255 & abs(R-G)>15/255 & R> G & R > B &...
reshape((max(mat_I)- min(mat_I)) >15/255, [s1,s2]);
index = uint8(index);
skin_img = repmat(index,[1,1,3]).*I;

  5 Comments

Show 2 older comments
darova
darova on 10 Sep 2019
Can i use your code for my study research?
Not for commercial purposes
Muhammad Farhan  Mughal
Muhammad Farhan Mughal on 10 Sep 2019
Yes, you can I also wrote that code for research purposes.

Sign in to comment.

More Answers (1)

darova
darova on 9 Sep 2019
See attached script
clc, clear
A = imread('masson.jpg');
subplot(1,2,1)
imshow(A)
RGB = [40 120 40]; % color you want
e = 40; % color shift / deviation
B = pix_in(A,RGB,e);
B = B + 255.*uint8(~B); % choosing white background
subplot(1,2,2)
imshow(B)

  7 Comments

Show 4 older comments

Sign in to comment.