Image Sharpening

Updated 25 Sep 2026

•
What it is
- •
  enhancing local contrast associated with edges and fine detail
•
Unsharp masking
- •
  obtain a smoothed image $\bar f$
- •
  form a detail mask $m=f-\bar f$
- •
  compute $g=f+k,m$ , with $k>0$
•
Tradeoffs
- •
  can amplify noise and produce halos
- •
  denoising before sharpening can help when noise is substantial
•
Implementation caution
- •
  compute using a type that supports intermediate negative or out-of-range values, then clip or rescale appropriately
•
Distinguish from
- •
  Edge Detection - identifies boundaries rather than primarily producing and enhanced-looking image
•
Related
- •
  Spatial Filtering

Linked references 4