Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!stanford.edu!leland.Stanford.EDU!hanauma!rick
From: rick@hanauma.stanford.edu (Richard Ottolini)
Newsgroups: comp.graphics
Subject: Re: contrast enhancer
Message-ID: <1991Apr8.155021.18518@leland.Stanford.EDU>
Date: 8 Apr 91 15:50:21 GMT
References: <1991Apr08.100539.13053@cs.ruu.nl>
Sender: news@leland.Stanford.EDU (Mr News)
Organization: Stanford University, Department of Geophysics
Lines: 16

In article <1991Apr08.100539.13053@cs.ruu.nl> clldomps@cs.ruu.nl (Louis van Dompselaar) writes:
>I'm looking for a reference to an article about how to
>enhance the contrast of computer pictures.

Some techniques:
(1) gamma correction: raise each sample to a signed power.  .5 or square root
is useful.  If you do this many times, then use table lookup for speed.

(2) Histogram equalization: sort the image samples into amplitude bins, then
redistribute them to fill the color space.

(3) Edge detector and derivative operators: 2-D FFT the image and weight the
higher frequencies.  A weight porportional to freuqency is the the first
derivative, but this may enchance contrast too strongly, so try something less.
A derivative operation can sometimes be implemented more cheaply by convolving
the image with small grid operater, e.g. {{1 -1} {-1 1}}.