i'm trying to enhanced image by gaussian high pass filter but it can't work... can you help me to repair my program? thangkyou before

33 views (last 30 days)
elfrida on 16 Aug 2013
% function GaussianHighPass [m n]=size(image.jpg); f_transform=fft2(image.jpg); f_shift=fftshift(f_transform); x0=m/2; y0=n/2; D0=70; a=12.53; b=-4; for x=1:m; for y=1:n; D=sqrt((x-x0)^2+(y-y0)^2); hpf(x,y)=a*(1-exp(-(D^2)/(2*(D0^2))))+b; end end filter_apply=conv2(image.jpg,hpf,'same'); figure,imshow(filter_apply) image_orignal=ifftshift(filter_apply); image_filter_apply=abs(ifft2(image_orignal)); figure,imshow(image_filter_apply,[])

Jan on 16 Aug 2013
Edited: Jan on 16 Aug 2013
As you are using imshow I'm guessing you have the image processing toolbox installed. In that case you might as well use the imfilter() function. The function
fspecial('Gaussian', hsize, sigma )
will generate a Gaussian lowpass filter.
High pass filtering can be achieved by subtracting the low pass filtered from the original image, e.g.
I = imread( 'peppers.png' );
I_filtered = I - imfilter( I, fspecial('Gaussian', 5, 1 ) );
imtool( I_filtered );
If you want to do the filtering explicitly, create the convolution matrix, e.g.
hp_filter_matrix = fspecial('Gaussian', 5, 1);
This gives you a 5 x 5 double matrix that you can convolve with your image to low pass filter it. As before, the high pass filtered representation can be obtained by subtraction.
2 CommentsShowHide 1 older comment
Jan on 16 Aug 2013
You are welcome! If that helped, you might want to mark it as 'Accepted Answer'.