Spatial Filtering

Updated 25 Sep 2026

•
What it is
- •
  computing output pixel values using a neighborhood in the image's spatial domain
•
Linear form
- •
  a local weighted sum: $g(x,y)=\sum_i\sum_j w(i,j)f(x-i,y-j)$
•
Nonlinear example
- •
  Median Filter selects an order statistic rather than computing a weighted sum
•
Choices
- •
  neighborhood size
- •
  weights or nonlinear rule
- •
  border handling
- •
  output numeric type and range
•
Applications
- •
  smoothing, denoising, sharpening, and edge detection
•
Caution
- •
  larger neighborhoods can suppress noise whil also removing useful detail
•
Related
- •
  Image Sharpening
- •
  Edge Detection
•
•

Linked references 7