Author Topic: Background Neutralization  (Read 401 times)

Bob Hertel

• Newbie
• Posts: 19
Background Neutralization
« on: 2012 July 07 04:08:26 »
If for each color channel, I were to use the HistogramTransformation tool, and manually set the Midtones value to the median value for that channel, is the result equivalent to performing Background Neutralization?

Philip de Louraille

• Posts: 253
Re: Background Neutralization
« Reply #1 on: 2012 July 07 04:14:40 »
I don't think so. Background means just that: background sky. E.g.: no nebulae, galaxies, ... Just stars and "empty" space. Background neutralization makes that sky neutral in color and may remove color biases in the picture.
Philip de Louraille

Cosmick

• Newbie
• Posts: 3
Re: Background Neutralization
« Reply #2 on: 2012 July 07 08:46:50 »
If for each color channel, I were to use the HistogramTransformation tool, and manually set the Midtones value to the median value for that channel, is the result equivalent to performing Background Neutralization?

I think this is closer to how Colour Calibration is done. Background Neutralization is a bit like setting the blackpoint of each channel to the same point on the histogram. As Philip said -
Quote
Background means just that: background sky. E.g.: no nebulae, galaxies, ... Just stars and "empty" space. Background neutralization makes that sky neutral in color and may remove color biases in the picture.
« Last Edit: 2012 July 07 08:55:03 by Cosmick »
Clear Skies

Mick

Juan Conejero

• PTeam Member
• PixInsight Jedi Grand Master
• Posts: 3895
Re: Background Neutralization
« Reply #3 on: 2012 July 07 15:59:01 »
Hi Bob,

Quote
If for each color channel, I were to use the HistogramTransformation tool, and manually set the Midtones value to the median value for that channel, is the result equivalent to performing Background Neutralization?

Not at all. The background neutralization algorithm is as follows:

INR = IR - BIR + b
ING = IG - BIG + b
INB = IB - BIB + b

where INR, ING and INB are the color channels of the resulting (background-neutralized image); IR, IG and IB are the original color channels; BIR, BIG and BIB are the mean backgrounds measured on the original channels, and b is the target mean background. Typically, b is defined by:

b = min( BRR, BRG, BRB )

where the BR's are the mean backgrounds measured on the background reference image.

Each mean background is estimated as the median of the pixels in the background sampling range (defined by the lower and upper limit parameters in the BackgroundNeutralization tool) for the corresponding channel of the original or background reference image.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/