11-02-2013, 03:14 PM
Hue-preserving Color Image Enhancement Without Gamut Problem
1Hue-preserving.pdf (Size: 428.14 KB / Downloads: 44)
Abstract
This paper discusses a general technique for extending the various grey-scale
transform for enhancement to color images while preserving the Hue of the im-
age [1]. The author do not talk about any particular transform, but instead provide
an innovative system using which we can extend almost any grey-scale image en-
hancing transform to color images. Using their new algorithm, one need not even
change color spaces while trying to work on an image, thus saving an enormous
amount of computation. However, they also agree that they talk in length only
about one way of generalizing those transforms (and generalize only Histogram
equalization) while in reality there exist many such extensions. The work on nd-
ing which of these transforms would be best for which kind of image is left as an
future research exercise. Also, they subsume a very important paper [2] which
forms a very integral part of the conclusions that they draw. Hence, while trying
to present their views on the subject, it would benet us greatly if we dig a deeper
into their assumptions and see why they are justied. A light introduction to what
their assumption was is also included in this term paper.
Introduction
Image enhancement is used to improve the quality of an image for visual perception of human
beings or for a machine for easier analysis. The set of pixel values of one image is transformed to a
new set of pixel values so that the new image formed is visually pleasing and is also more suitable
for analysis. We are well acquainted with the main techniques for image enhancement for grey scale
images:
Contrast stretching,
Slicing,
Histogram equalization, .
The Algorithm
A nonlinear transformation used in contrast enhancement
for grey scale images. For example, S-type transformation is listed earlier in this section.
(lx), hence, may exceed 1. Now there will be such cases that value of x
0
k may exceed 1 and thus
resulting in gamut problem. A possible solution to this is to transform the color vector to CMY
space and process it there. This will be dealt with in two separate cases.
Histogram Equalization
Histogram equalization adjusts the histogram of the image to make it resemble to a uniform distribution.
It works extremely well for contrast enhancement of gray scale images. But, performing
the same operation on the RGB plane changes the hue of the pixel values of the image. So what
is sometimes done is to use histogram equalisation only on the luminance/saturation plane of the
image in the HSV color space. This enhances the image without aecting the hue of the pixels.