photo sharing and upload picture albums photo forums search pictures popular photos photography help login
Bob Walker | all galleries >> Picture Window Pro Tutorials >> Removing Hot Pixels >> Making a Hot Pixel Mask > Absolute Difference Transform
previous | next

Absolute Difference Transform

This is a screen shot of the Composite Transform, in Absolute Difference mode, used to calculate
the difference between the averaged dark frame and the twice median-blurred averaged dark frame.

The result of this transform is one of the three input options to the hot pixel threshold step.
This result has the best signal to noise ratio of the three options (the other two being
(a) one of the dark frames, or (b) the averaged dark frame).

Why the absolute difference operation instead of a subtract operation? Here, there's
no compelling reason, because all the hot pixels are brighter than their background.

Skip this paragraph if you do not care about dead pixels. The absolute difference is preferred
if your camera also has some dead pixels (always black). In this case, you could follow
very much the same procedure outlined here to collect all the dead pixels along with
the hot pixels. The only modification would be that instead of starting with a dark frame,
you would want to start with a "gray frame" -- a long, out-of-focus exposure of a
white background, adjusted so the image histogram of the gray frame is centered somewhere
near the middle of the range of brightness values in HSV space. In such an image, the brightest
pixels are hot, and the darkest ones are dead. The absolute difference transform would render the
result of this step with both hot pixels and dead pixels as brighter than their background.

My camera does not have any dead pixels, so I cannot illustrate that enhancement to this technique.


other sizes: small medium original auto
comment