Most of the time, the obtained digital images are either coded or
they are subjected to modifications for different purposes. The Image
Quality Metric is used when one wishes to know how these modifications
distort the image. While determining the level of distortion of the
image, it is of paramount importance to take into consideration such
questions as to whom/to what/for what purpose. An image can be
solely used at visual examination and it can also be used for the medical
diagnosis or various analysis. The level of distortion of the image has
different implications for each one of the applications and it determines
whether the image is usable for the relevant application. For this
reason, the “application-specific” image quality measurement methods
are of great significance [1].
The IQMs in the literature are separated into 3 groups: fullreference
IQM [2], reduced-reference IQM [3] and no-reference IQM
[4]. The Full-reference IQM need both the distorted and original image.
By taking the original image as a reference, how different the distorted
image is calculated. However, there is not reference image as regards
the no-reference IQM. For this reason, while comparing, only one
image is taken into consideration. In the reduced-reference IQMs
which can be considered to be a method between these two methods,
the quality of the distorted image is measured by partially using the
reference image [5–7].
There are still serious shortcomings regarding the development of
algorithms which take into account all characteristics of the HVS,
which are robust, practical and adequately respond to the needs. In the
IQM methods in the literature, it has been found that;
(a) Results have been produced for the modifications not perceived by
the HSV.
(b) The physiological characteristics of the HVS are not fully taken into
consideration.
(c) The HVS's sensitivity to the luminance, texture and edge (psychophysiological
characteristics) are not taken into consideration.
(d) Even though the color and luminance information is paramount
importance for the HVS (physiological properties), the quality
measurement procedures are carried out after the conversion of
the images into the grayscale image [7–12].
The basic objective in this paper is to develop a new quality metric
based on the HVS for the digital images and obtain a metric producing
the numeric quality results compatible with the HVS by taking the
image as a whole.