extracting the channels from the corrected image

iam working with rgb fundus images

my task is to convert the rgb image to hsv space and extract the channels and

perform brightness correction in the image

now i am little confused on how to extract the channels of the corrected image

the corrected image is not saved as jpg format

could anyone help me to solve this issue

i have also attached the coding

thank you in advance

Answer 1

[Ameer Hamza]

You converted the image to RGB after applying correction. You can extract the RGB channels by indexing along third dimension

R = newRGBImage(:,:,1);
G = newRGBImage(:,:,2);
B = newRGBImage(:,:,3);

[LAXMI PRIYANKA]

iam getting the folowing error

Undefined function 'newRGBImage' for input arguments of type 'char'.

[Ameer Hamza]

LAXMI, the newRGBImage is defined in your code. Add this after these lines in your code

meanV = mean2(V);
newV = meanV + 1.5 * (V - meanV); % Increase contrast by factor of 1.5
newS = 1.5 * S; % Increase contrast by factor of 1.5
newHSVImage = cat(3, H, newS, newV);
newRGBImage = hsv2rgb(newHSVImage);

R = newRGBImage(:,:,1);
G = newRGBImage(:,:,2);
B = newRGBImage(:,:,3);

[LAXMI PRIYANKA]

it worked great thank you sir

i would also like to ask you another help regarding applying CLAHE to the extraxcted green channel

i have been searching for weeks together and would be really happy if you could hep me with the source code

thanking you in advance

[Ameer Hamza]

What about this: https://www.mathworks.com/help/images/ref/adapthisteq.html

[LAXMI PRIYANKA]

i read it sir

when i use clahe i get the following error

Undefined function 'localcontrast' for input arguments of type 'double'.

i have also attached the codings

[Ameer Hamza]

Change the line 40 and 41 in your code to

data = localcontrast(im2uint8(newRGBImage), edgeThreshold, amount);
PSNR1 = psnr(im2double(data), newRGBImage);

[LAXMI PRIYANKA]

sir just now tried those codins with rgb image in the place of tif image

not working

getting errors

Error using adapthisteq

Expected input number 1, I, to be two-dimensional.

Error in adapthisteq>parseInputs (line 416)

validateattributes(I, {'uint8', 'uint16', 'double', 'int16', 'single'}, ...

Error in adapthisteq (line 154)

[I, selectedRange, fullRange, numTiles, dimTile, clipLimit, numBins, ...

Error in s (line 2)

J = adapthisteq(I,'clipLimit',0.02,'Distribution','rayleigh');

Error using adapthisteq

Expected input number 1, I, to be two-dimensional.

Error in adapthisteq>parseInputs (line 416)

validateattributes(I, {'uint8', 'uint16', 'double', 'int16', 'single'}, ...

Error in adapthisteq (line 154)

[I, selectedRange, fullRange, numTiles, dimTile, clipLimit, numBins, ...

Undefined function or variable 'L'.

Undefined function or variable 'L'.

Undefined function 'rgb2lab' for input arguments of type 'double'.

[LAXMI PRIYANKA]

i tried with the coding in line 40 and 41

error displayed as

Undefined function 'localcontrast' for input arguments of type 'uint8'.

[LAXMI PRIYANKA]

will this be due to the version of matlab?

i am currently using matlab2014a

if it is that i have to install the latest or the next version say2015 or higher

what should i do at this time

help me please with this regard

[Ameer Hamza]

Can you share your input image?

[Ameer Hamza]

I am not sure if this is related to MATLAB version. I am running the code on R2020a.

[LAXMI PRIYANKA]

herewith iam attaching my image

[Ameer Hamza]

The code works fine on R2020a using this image, but there is any error when making the final figures. There is a mistake in typing the name of variables. There is no variable name image in your code, but these lines are trying to call imshow on it.

figure('name','LCM-CLAHE','NumberTitle','off')
%da=adapthisteq(data);
%%%% <------ There is no variable named image in your code
imshow([image uint8(output)]);          % Concatenate original and CLAHE image

figure('name','Histogram of input image','NumberTitle','off'); imhist(image);               %Show the histogram of original image, LCM output and CLAHE output
figure('name','Histogram of LCM output','NumberTitle','off');imhist(data);
figure('name','Histogram of CLAHE output','NumberTitle','off');imhist(uint8(output));
PSNR = psnr(uint8(output), image);
fprintf('\n The Peak-SNR value of LCM-CLAHE is %0.4f', PSNR);
normal = adapthisteq(image,'clipLimit',0.02,'Distribution','rayleigh');
%imshow([image normal]);
PSNR0 = psnr(normal,image);
fprintf('\n The Peak-SNR value of CLAHE is %0.4f', PSNR0);
he = histeq(image);
PSNR2 = psnr(he,image);
fprintf('\n The Peak-SNR value of HE is %0.4f', PSNR2);

[LAXMI PRIYANKA]

so i am supposed to use newRGBImage in the place of image .

am i right sir

[LAXMI PRIYANKA]

sir could u help me to solve the error

what should i specify in the place of image in the coding

[Ameer Hamza]

I am not sure. I also though this, but output just have one channel and newRGBImage has three channel so the lines

imshow([newRGBImage uint8(output)]); 
..
..
PSNR0 = psnr(normal,newRGBImage);

give error.

[LAXMI PRIYANKA]

sir 
i have to apply clahe for the green channel extracted from the newRGBImage 
so now i'll try with the G channel , but the G channel is of  data type 
double ie  960x1440 double 
what should i do please help me 

[LAXMI PRIYANKA]

data = localcontrast(im2int8(G), edgeThreshold, amount);

PSNR1 = psnr(im2double(data), G);

Error: Undefined function 'im2int8' for input arguments of type 'double'.

Undefined function or variable 'data'.

[Ameer Hamza]

The correct function name is im2uint8, not im2int8

data = localcontrast(im2uint8(G), edgeThreshold, amount);

[LAXMI PRIYANKA]

Undefined function 'localcontrast' for input arguments of type 'uint8'. again i get the same error

[Ameer Hamza]

What is the complete error message?

[LAXMI PRIYANKA]

Warning: Image is too big to fit on screen; displaying at 50%

> In imuitools\private\initSize at 71

In imshow at 282

In maincodecheck at 25

Warning: Image is too big to fit on screen; displaying at 50%

> In imuitools\private\initSize at 71

In imshow at 282

In maincodecheck at 31

Warning: Image is too big to fit on screen; displaying at 50%

> In imuitools\private\initSize at 71

In imshow at 282

In maincodecheck at 33

Undefined function 'localcontrast' for input arguments of type 'uint8'.

Error in maincodecheck (line 40)

data = localcontrast(im2uint8(G), edgeThreshold, amount);

these are the errors which i get while running the coding as a whole

[Ameer Hamza]

Rest are errors. The actual warning is

Undefined function 'localcontrast' for input arguments of type 'uint8'.

Which is probably caused by the old version of MATLAB. I am not sure which input type is compatible with localcontrast in this version of MATLAB.

[LAXMI PRIYANKA]

ok sir thank you for your help

and is it possible to guide me so as to get the udated version of Matlab

i am also trying with other codings and i'll ask you after executing the codes

[Ameer Hamza]

If you don't already have one, you will need to get a MATLAB license to use the latest release.

[LAXMI PRIYANKA]

does it cost high

[Ameer Hamza]

Depend on your requirement and the license type applicable to you. get more information here: https://www.mathworks.com/pricing-licensing.html

[LAXMI PRIYANKA]

thank you sir

[LAXMI PRIYANKA]

sir now i tried with adaphist function for applying clahe

[f_name p_name]=uigetfile('*','select the image');

input_img=imresize(imread(strcat(p_name,f_name)),[512,512]);

imshow(input_img),title('Input Image');

[xRes,yRes]=size(input_img);

Min=0;

Max=255;

NrX=2;

NrY=2;

NrBins=256;

Cliplimit=0.01;

% I = imread('tire.tif');

J = adapthisteq(im2uint8(input_img),'clipLimit',0.02,'Distribution','rayleigh');

imshowpair(im2uint8(input_img),J,'montage');

title('Original Image (left) and Contrast Enhanced Image (right)')

the abve codings after execution displays the following error

Error using adapthisteq

Expected input number 1, I, to be two-dimensional.

Error in adapthisteq>parseInputs (line 416)

validateattributes(I, {'uint8', 'uint16', 'double', 'int16', 'single'}, ...

Error in adapthisteq (line 154)

[I, selectedRange, fullRange, numTiles, dimTile, clipLimit, numBins, ...