Image Processing Reference
In-Depth Information
Although QCA and CA as architectures share some common characteristics they
are mostly different as it is described later in this chapter. A basic CA cell implemen-
tation using QCA has already been proposed at the literature [33]. The implemented
QCA circuit is presenting much better characteristics compared to its conventional
VLSI implementation in terms of circuit area, clock frequency and power consump-
tion. According to these results it can be easily assumed that all CA systems which
have been designed and implemented using conventional VLSI technology now can
be implemented using QCA [32, 55]. These new QCA circuits will operate near 1
THz clock frequencies, they will cover 100 times (or more) less area than their previ-
ous implementations and they will present extremely low power consumptions [29].
In the following section, basic concepts of mathematical morphology will be pre-
sented, and a simple description of the quantum-dot cellular automata will follow in
Sect. 4.3. The methodology that should be followed for a correct and efficient design
of a QCA circuit will be given in Sect. 4.4. The QCA design of the morphological
erosion and dilation operations will be presented in Sect. 4.5 as implementation
examples. Finally, conclusions will be drawn in Sect. 4.6.
4.2
Mathematical Morphology
In recent years, the evolution of digital recording technology has made digital image
and video applications an important aspect of daily life. But, in many cases and due
to a variety of reasons, the quality of digital images and videos reduces to unsat-
isfactory levels. Common reasons for this corruption are the unconstrained camera
motion, the uncontrolled indoor or outdoor environments, the existence of clutter
and the exposure to noise during signal transmission. There are many algorithms
proposed in the literature trying to improve the quality of corrupted images. Such al-
gorithms are usually implemented in hardware in order to improve the performance
of image and video processing systems and satisfy the real time processing require-
ments desired in many cases. In such systems, mathematical morphology operators
and filters are extensively used as parts of more complex algorithms.
Mathematical morphology is a field of image processing that offers a uni-
fied approach and powerful tools applied to image and video processing applica-
tions [13, 51, 52], including thinning, thickening, skeletonising, shape extraction,
segmentation, object detection and tracking, noise reduction, feature point selec-
tion, face recognition and verification and many more [14, 31, 36, 56]. There are
also some more complex morphological algorithms such as interpolation of 3D bi-
nary objects [6], robust 2D and 3D object representation [5]. In other cases, mor-
phological filters are applied as parts of more complicated algorithms in order to
achieve better performance [21, 46].
The basic operations of mathematical morphology are: erosion, dilation, opening
and closing. They are explicitly defined, and due to their simplicity they are used
in many image analysis and computer vision problems. All basic operations take
two sets of data as input: the input image and the structuring element. The simplest
version of mathematical morphology is applied on binary images and uses binary
Search WWH ::
Custom Search