Filtering

 (i) Write an m-file "mymedian.m" which will take an image and filter size and perform median filtering. Display input and output image. Use 

help median

function imm=mymedian(imn,msize)
[h w]=size(imn);
kernel_size=?
imm=imn;
for i=msize+1:(h-msize)
  for j=msize+1:(w-msize)
    imm(i,j)=median(reshape(imn(i-msize:i+msize,j-msize:j+msize),1,kernel_size));
  end
end

What is the median value of an image?

Comment on the complexity (number of operations per pixel) of the algorithm. 

(ii) Write an m-file "mygaussian.m" which will take an image and filter size and perform Gaussian filtering. Display input and output image.

function img=mygaussian(imn,sigma)
g=gauss(sigma);

img=?(iii) Perform median and Gaussian filtering on  face_g.png, face_m.png,statue_m.png.
What kind of noise the median filer is useful for?
What kind of noise the Gaussian filer is useful for?