Image Sharpening

In image sharpening the goal is to highlight fine details in an image. That is, to enhance details that have been blurred. Fine details in the frequency domain correspond to high frequencies, thus the use of high-pass filters for image sharpening. In enhancing fine details the price to pay is also the enhancement of noise.

In the spatial domain high-pass filters greater than 3x3 are seldom needed. An example of the filters used is a Laplacian. The origin or "sweet spot" of the filter is its center.

#
# Laplacian High-pass filter
#
 0.0 -1.0  0.0
-1.0  4.0 -1.0
 0.0 -1.0  0.0

Some characteristics of high-pass filters are:

• positive coefficients near the center
• negative coefficients in the outer periphery
• sum of filter coefficients is zero (no DC term)

Before processing it is recommended to clip negative pixel values, and not to take absolute values because large negative values introduce spikes in the image.

Experiment: high filtering emphasis

• Building a visual program (workspace) in cantata
• Execute the visual program laplacian-sharp.wk