- Forming a new image whose pixel values are transformed from original pixel values
- See system (cv)
moving average
- linear shift invariant (LSI)
- for convolution
- represent k,l as n-k and m-l BECAUSE
- we shift the flipped kernel across the image to center the kernel over the corresponding input image pixel
- multiply values of input image in neighborhood of (n,m) with corresponding values of flipped kernel
- sum products
- (just legacy from pure math)
image segmentation
