how can i make such an image by apllying DC processing?
조회 수: 1(최근 30일)
input image is a 256X256 gray image, and I used dct2 to make a DCT image. then professor said I should set DC 32 X 32 =0 and IDCT back to get the output image,
and i dont know how to set 32X32 to 0....
please tell me the way to set 32x32 part of the dct input image to 0.
Image Analyst 2019년 11월 23일
Regarding your new, edited question, just figure out the rows and columns to set to zero. If you use fft2(), the origin is at the 4 corners. You'd have to erase each corner one at a time. Or else use fftshift() to move the origin to the middle, then erase, and then use ifftshift() to shift the origin back to the corners before caling ifft2. Here's a start
[rows, columns] = size(grayImage);
ft = fft2(grayImage);
ft = fftshift(ft);
row1 = rows/2 - 15;
row2 = row2 + 31;
col1 = ....
col2 = ....
ft(row1:row2, col1:col2) = 0;
ft = ifftshift(ft);
grayImage = real(ifft2(ft));
Image Analyst 2019년 11월 22일
Yes, but not if you erase the central 32x32 chunk of it, representing the low spatial frequencies. It will have the effect of a high pass filter since you're setting all the low frequencies to zero. You will see enhanced edges, or even mostly all edges, in the output.