Gray Scale to Pseudo color transformation

   Transformation of a gray scale image into pseudo color image helps in better visualization of the image. In this tutorial, different ways to apply pseudo color transformation to a gray scale image will be discussed along with the MATLAB Code.

The main idea behind pseudo color transformation is to perform three independent transformation (RED,GREEN and BLUE) on the grayscale or intensity image and map  the corresponding intensity value in the image to the result obtained.

Steps to be performed:

Functional Block Diagram

Reference: Digitial Image Processing by Gonzalez

MATLAB CODE:

%READ AN INPUT IMAGE

A = imread('cameraman.tif');

%PRE-ALLOCATE A MATRIX

Output = zeros([size(A,1) size(A,2) 3]);

%Define a colormap

map = colormap(jet(256));

%Assign the columns to 1-D RED,GREEN and BLUE

Red = map(:,1);

Green = map(:,2);

Blue = map(:,3);

%MAP THE COLORS BASED ON THE INTENSITY OF THE IMAGE

Output(:,:,1) = Red(A);

Output(:,:,2) = Green(A);

Output(:,:,3) = Blue(A);

Output = im2uint8(Output);

%DISPLAY THE IMAGE

imshow(Output);

%Save the image in PNG or JPEG format

imwrite(Output,'pseudo_color.jpg');

 Checkerboard with colormaps:

%READ AN INPUT IMAGE

A = imread('coins.png');

%RESIZE THE IMAGE

A = imresize(A,[256 256]);

%PRE-ALLOCATE THE OUTPUT MATRIX

Output = zeros([size(A,1) size(A,2) 3]);

%DEFINE SOME COLORMAPS

maps={'jet(256)';'hsv(256)';'cool(256)';'spring(256)';'summer(256)';'parula(256)';'hot(256)'};

%COLORMAP 1

map=colormap(char(maps(7)));

Red = map(:,1);

Green = map(:,2);

Blue = map(:,3);

R1 = Red(A);

G1 = Green(A);

B1 = Blue(A);

%COLORMAP 2

map=colormap(char(maps(6)));

Red = map(:,1);

Green = map(:,2);

Blue = map(:,3);

R2 = Red(A);

G2 = Green(A);

B2 = Blue(A);

%SIZE OF THE MATRIX

sz=64;

ind=0;

for i=1:sz:size(A,1)-(sz-1)

for j=1:sz:size(A,2)-(sz-1)

%APPLY COLORMAPS BASED ON THE SIZE SZ

if(mod(ind,2)==0)

            Output(i:i+(sz-1),j:j+(sz-1),1) = R1(i:i+(sz-1),j:j+(sz-1));

            Output(i:i+(sz-1),j:j+(sz-1),2) = G1(i:i+(sz-1),j:j+(sz-1));

            Output(i:i+(sz-1),j:j+(sz-1),3) = B1(i:i+(sz-1),j:j+(sz-1));

else

            Output(i:i+(sz-1),j:j+(sz-1),1) = R2(i:i+(sz-1),j:j+(sz-1));

            Output(i:i+(sz-1),j:j+(sz-1),2) = G2(i:i+(sz-1),j:j+(sz-1));

            Output(i:i+(sz-1),j:j+(sz-1),3) = B2(i:i+(sz-1),j:j+(sz-1));

end

        ind=ind+1;

end

    ind=ind+1;

end

Output = im2uint8(Output);

%FINAL IMAGE

imshow(Output);

EXPLANATION:

The above code is to combine two colormaps to obtain checkboard effect on the image. To obtain checkboards in different size change the 'sz' variable with 4,8,16..etc.

UPPER OR LOWER TRIANGLE :

MATLAB CODE:

%READ INPUT IMAGE

A = imread('coins.png');

%RESIZE IMAGE

A = imresize(A,[256 256]);

%PRE-ALLOCATE THE OUTPUT MATRIX

Output = ones([size(A,1) size(A,2) 3]);

%COLORMAPS

maps={'jet(256)';'hsv(256)';'cool(256)';'spring(256)';'summer(256)';'parula(256)';'hot(256)'};

%COLORMAP 1

map=colormap(char(maps(1)));

Red = map(:,1);

Green = map(:,2);

Blue = map(:,3);

R1 = Red(A);

G1 = Green(A);

B1 = Blue(A);

%COLORMAP 2

map=colormap(char(maps(7)));

Red = map(:,1);

Green = map(:,2);

Blue = map(:,3);

%RETRIEVE POSITION OF UPPER TRIANGLE

[x,y]=find(triu(Output)==1);

Output(:,:,1) = Red(A);

Output(:,:,2) = Green(A);

Output(:,:,3) = Blue(A);

for i=1:numel(x)

        Output(x(i),y(i),1)=R1(x(i),y(i));

        Output(x(i),y(i),2)=G1(x(i),y(i));

        Output(x(i),y(i),3)=B1(x(i),y(i));

end

Output = im2uint8(Output);

%FINAL IMAGE

imshow(Output);

EXPLANATION:

Upper triangular part of the image matrix with a colormap and lower triangular part of the image matrix with different colormap.

Image Rotation in MATLAB - Examples without imrotate function – Part 2

To flip an image left to right without using ‘fliplr’

Grayscale:

 A = imread('cameraman.tif');

 LR = A(1:end,end:-1:1);

 figure, subplot(1,2,1);imshow(A);title('original');         

         subplot(1,2,2);imshow(LR);title('Flipped left to right');

RGB Image:

RGB = imread('peppers.png');

 LR  = RGB(1:end,end:-1:1,:);

 figure, subplot(1,2,1);imshow(RGB);title('original');         

         subplot(1,2,2);imshow(LR);title('Flipped left to right');

To flip an image upside down without using ‘flipud’

Gray Scale:

 A = imread('liftingbody.png');

 UD = A(end:-1:1,1:end);

 figure, subplot(1,2,1);imshow(A);title('original');         

         subplot(1,2,2);imshow(UD);title('Upside Down');

RGB Image:

 UD = RGB(end:-1:1,1:end,:);

 figure, subplot(1,2,1);imshow(RGB);title('original');         

         subplot(1,2,2);imshow(UD);title('Upside Down');

MATLAB code to rotate a gray scale Image:

A = imread('cameraman.tif');

%Specify the degree

deg = 35;

%Find the midpoint

if(deg > 155)

    midx = ceil((size(A,1))/2);

    midy = ceil((size(A,2))/2);

else

    midx = ceil((size(A,2))/2);

    midy = ceil((size(A,1))/2);

end

[y,x] = meshgrid(1:size(A,2), 1:size(A,1));

[t,r] = cart2pol(x-midx,y-midy);

t1 = radtodeg(t)+deg;

%Convert from degree to radians

t = degtorad(t1);

%Convert to Cartesian Co-ordinates

[x,y] = pol2cart(t,r);

%Add the mid points to the new co-ordinates

tempx = round(x+midx);

tempy = round(y+midy);

if ( min(tempx(:)) < 0 )

   

newx = max(tempx(:))+abs(min(tempx(:)))+1;

tempx = tempx+abs(min(tempx(:)))+1;

else

    newx = max(tempx(:));

end

if( min(tempy( : )) < 0 )

   

newy = max(tempy(:))+abs(min(tempy(:)))+1;

tempy = tempy+abs(min(tempy(:)))+1;

else

    newy = max(tempy(:));

end

tempy(tempy==0) = 1;

tempx(tempx==0) = 1;

C = uint8(zeros([newx newy]));

for i = 1:size(A,1)

    for j = 1:size(A,2)

        C(tempx(i,j),tempy(i,j)) = A(i,j);

       

    end

  

end

figure,imshow(C);

Output = C;

%FILL THE HOLES OR GAPS-NEAREST NEIGHBOR

for i = 2:size(C,1)-1

    for j = 2:size(C,2)-1

       

        temp = C(i-1:i+1,j-1:j+1);

        if(temp(5)==0&&sum(temp(:))~=0)

            pt = find(temp~=0);

           

            [~,pos] = sort(abs(pt-5));

            Output(i,j) = temp(pt(pos(1)));

          

        end

       

    end

end

figure,imshow(uint8(Output));

EXPLANATION:

This is a faster implementation of the older version and also the gaps are filled with the nearest neighbor .

MATLAB code to rotate a RGB Image

A=imread('peppers.png');

%Specify the degree

deg=45;

%Find the midpoint

if(deg>155)

    midx = ceil((size(A,1))/2);

    midy = ceil((size(A,2))/2);

else

    midx = ceil((size(A,2))/2);

    midy = ceil((size(A,1))/2);

end

[y,x] = meshgrid(1:size(A,2), 1:size(A,1));

[t,r] = cart2pol(x-midx,y-midy);

t1 = radtodeg(t)+deg;

%Convert from degree to radians

t = degtorad(t1);

%Convert to Cartesian Co-ordinates

[x,y] = pol2cart(t,r);

tempx = round(x+midx);

tempy = round(y+midy);

if ( min(tempx ( : ) ) < 0 )

   

newx = max(tempx(:))+abs(min(tempx(:)))+1;

tempx = tempx+abs(min(tempx(:)))+1;

else

    newx = max(tempx(:));

end

if( min(tempy ( : ) ) < 0 )

   

newy = max(tempy(:)) + abs(min(tempy(:)))+1;

tempy = tempy + abs(min(tempy(:)))+1;

else

    newy = max(tempy(:));

end

tempy(tempy==0) = 1;

tempx(tempx==0) = 1;

C = uint8(zeros([newx newy 3]));

for i = 1:size(A,1)

    for j = 1:size(A,2)

        

        C( tempx(i,j),tempy(i,j) ,:) = A(i,j,:);

      

       

    end

  

end

figure,imshow(C);

Output=C;

%FILL THE HOLES OR GAPS - NEAREST NEIGHBOR

for i = 2:size(C,1)-1

    for j = 2:size(C,2)-1

       

        temp = C(i-1:i+1,j-1:j+1,:);

        if(temp(5)==0 && sum(temp(:))~=0)

           

             pt = find(temp(:,:,1)~=0);

            [~,pos] = sort(abs(pt-5));

           

            temp = C(i-1:i+1,j-1:j+1,1);

            Output(i,j,1) = temp(pt(pos(1)));

           

            temp = C(i-1:i+1,j-1:j+1,2);

            Output(i,j,2) = temp(pt(pos(1)));

           

            temp = C(i-1:i+1,j-1:j+1,3);

            Output(i,j,3) = temp(pt(pos(1)));

        end

       

    end

end

figure,imshow(uint8(Output));title([num2str(deg),'Degree']);