Comparative Analysis of Spatial Filters for Image Rectification and Computational Complexity
Date
2008
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Department of Computer Science and Information Technology
Abstract
Image Rectification is an important preprocessing step in many fields such as computer
vision, pattern matching and many real time image-processing applications. Rectifying
different things is image Rectification. But here the algorithm that helps to rectify
impurity presents in image is proposed. This impurity can be noise and distortions. The
sources of impurity in digital images arise during image acquisition (digitization) and /or
transmission. Impurity has two kinds of properties 1) Spatial properties 2) Frequency
properties. Frequency properties refer to the frequency content of impurity in the Fourier
sense. The spatial characteristics are concerned with statistical behavior of the gray level
values in the impurity component. So the proposed algorithm focuses in spatial properties
of impurity. There are two types of spatial filters. But noise reduction can be achieved
effectively with a nonlinear filter whose basic function is to compute the median graylevel
value in the neighborhood in
which the filter is located.
There
exist many
works on minimizing
impurity
of
images
by using spatial filters. The
most
widely
used spatial
filters
are Median Filter,
Kuwahara filter
and Gaussian filter.
All
are responsible to minimize
impurity to some
extent.
The proposed algorithm that
helps
to minimize
impurity
is better than these filters mentioned
above. Here, checking of
the
minimization
of error is done by calculating
RMS errors between input image
and
output
image.
The proposed algorithm
is also based
on the spatial filters. Experiment
showed
that the final rectified images
are satisfactory
and it makes
it easy for further
image
manipulation.
The
efficiency of an algorithm
depends on the
complexity
(time
and space). Time
complexity
is total time
required for execution, whreas
space complexity
is total memory
space
required for execution
of an algorithm.
Complexity
of Median Filter,
Kuwahara
filter
and Gaussian filter and proposed filter
are analysed for better efficiency.
Keywords:
Description
Keywords
Noise, Distortions, Acquisition, Spatial, Impurity, Illumination, Gray level, Filter mask, Root Mean Square Error, Pixel