An Efficient Cum Improved Resolution Enhancement Using Zero Padding and Curvelet Transforms

advertisement
International Journal of Engineering Trends and Technology (IJETT) – Volume 15 Number 1 – Sep 2014
An Efficient Cum Improved Resolution
Enhancement Using Zero Padding and
Curvelet Transforms
1
1
Tirveedi Prasad, 2 Ch.Pullarao M.Tech
systems &signal processing, L.B.R.C.E, Mylavaram.
2
M.tech, ,E.C.E, Mylavaram
ABSTRACT: The main objective of this
wavelet transforms used in image processing. DWT
thesis is to design an enhanced and efficient
decomposes an image into different subband images,
resolution enhancement technique using zero padding
namely low-low (LL), low- high (LH), high-low
and curvelet transforms. In the existing paper the
(HL), and high-high (HH). Another recent wavelet
authors proposed a combination of SWT and DWT
transform which has been used in several image
for edge refinement and to improve lossless
processing applica- tions is stationary wavelet
resolution which is stated by the PSNR values as a
transform (SWT) [9]. In short, SWT is sim- ilar to
development in the edge detection and calculating the
DWT but it does not use down-sampling hence the
PSNR values CURVELETS were proposed in this
subbands will have the same size as the input image.
paper. So, through all the transformations a high
While inspired by the peculiar problem of such
PSNR value is obtained and spatial frequency, MSE
methods, the proposed approach also works for other
were also calculated and stated in the table along with
image enhancement methods that either introduce.
the results.
The interpolated high frequency subbands and the
SWT high fre-quency subbands have the same size
KEYWORDS:
PSNR,
CURVELETS,
ZERO
PADDING, DWT, MSE
which means they can be added with each other. The
new corrected high frequency subbands can be
interpolated further for higher enlargement. Also it is
INTRODUCTION:
ALTHOUGH the field of
image enhancement has been active since before
digital imagery achieved a consumer status, it has
never stopped evolving. The present work introduces a novel multi-resolution denoising method,
tailored to address a specific image quality problem
that
arises
when
using
image
enhancement
algorithms based on random spray sampling. Image
resolution enhancement in the wavelet domain is a
relatively new research topic and recently many new
known that in the wavelet domain, the low resolution
image is obtained by lowpass fil- tering of the high
resolution image [16]. In other words, low frequency
subband is the low resolution of the original image.
Therefore, instead of using low frequency subband,
which contains less information than the original high
resolution image, we are using the input image for
the interpolation of low frequency subband image.
Using input image instead of low frequency subband
increases the quality of the super image resolution.
algorithms have been pro- posed [4]–[7]. Discrete
wavelet transform (DWT) [8] is one of the re- cent
ISSN: 2231-5381
http://www.ijettjournal.org
Page 16
International Journal of Engineering Trends and Technology (IJETT) – Volume 15 Number 1 – Sep 2014
EXISTENCE
IMAGE
RESOLUTION
ENHANCEMENT:
frequency sub-band increases the quality of the super
resolved image. By interpolating input image byα/2,
In image resolution enhancement by using
and high frequency sub-bands by 2 and α in the
interpolation the main loss is on its high frequency
intermediate
components (i.e., edges), which is due to the
respectively, and then by applying IDWT, as
smoothing caused by interpolation. In order to
illustrated, the output image will contain sharper
increase the quality of the super resolved image,
edges than the interpolated image obtained by
preserving the edges is essential. In this work, DWT
interpolation of the input image directly. This is due
has been employed in order to preserve the high
to the fact that, the interpolation of isolated high
frequency components of the image. The redundancy
frequency components in high frequency sub-bands
and shift invariance of the DWT mean that DWT
and using the corrections obtained by adding high
coefficients are inherently interpolable .
frequency sub-bands of SWT of the input image, will
and
final
interpolation
stages
preserve more high frequency components after the
In this correspondence, one level DWT
interpolation than interpolating input image directly.
(with Daubechies 9/7 as wavelet function) is used to
decompose an input image into different sub-band
INTERPOLATION
images. Three high frequency sub-bands (LH, HL,
One of the commonly used techniques for
and HH) contain the high frequency components of
image resolution enhancement is Interpolation.
the input image. In the proposed technique, bi-cubic
Interpolation has been widely used in many image
interpolation with enlargement factor of 2 is applied
processing applications such as facial reconstruction,
to high frequency subband images. Down-sampling
multiple description coding, and super resolution.
in each of the DWT sub-bands causes information
There are three well known interpolation techniques,
loss in the respective sub-bands. That is why SWT is
namely nearest neighbor interpolation, bilinear
employed to minimize this loss.
interpolation, and bicubic interpolation. Interpolation
is the process of using known data values to estimate
The interpolated high frequency sub-bands
and the SWT high frequency sub-bands have the
unknown
data
values.
Various
interpolation
techniques are often used in the atmospheric sciences
same size which means they can be added with each
other. The new corrected high frequency sub-bands
Nearest neighbor interpolation:
can be interpolated further for higher enlargement.
Nearest-neighbor interpolation (also known as
Also it is known that in the wavelet domain, the low
proximal interpolation or, in some contexts, point
resolution image is obtained by lowpass filtering of
sampling) is a simple method of multivariate
the high resolution image. In other words, low
interpolation
frequency sub-band is the low resolution of the
Interpolation is the problem of approximating the
original image. Therefore, instead of using low
value for a non-given point in some space when
frequency sub-band, which contains less information
given some colors of points around (neighboring) that
than the original high resolution image, we are using
point. The nearest neighbor algorithm selects the
the input image for the interpolation of low frequency
value of the nearest point and does not consider the
sub-band image. Using input image instead of low
values of neighboring points at all, yielding a
ISSN: 2231-5381
in
http://www.ijettjournal.org
one
or
more
dimensions.
Page 17
International Journal of Engineering Trends and Technology (IJETT) – Volume 15 Number 1 – Sep 2014
piecewise-constant interpolation. The algorithm is
transformations. To implement the image resolution
very simple to implement and is commonly used
enhancement we adopt for zero padding fast curvelet
(usually along with mipmapping) in real-time 3D
transformations.
rendering to select color values for a textured surface.
proposed scheme is most elaborate and easy to say
The value indicates
that
the
most effective and efficient to work on. The results
Bilinear interpolation: In computer vision and
were discussed in the next section. Curvelet
image processing, bilinear interpolation is one of the
transformation to decompose a low resolution image
basic re-sampling techniques. In texture mapping, it
into different sub-bands. Then the three high
is also known as bilinear filtering or bilinear texture
frequency sub-band images have been interpolated
mapping, and it can be used to produce a reasonably
using bi-cubic interpolation. The high frequency sub-
realistic image. An algorithm is used to map a screen
bands obtained by dct of the input image are being
pixel location to a corresponding point on the texture
incremented into the interpolated high frequency sub-
map. A weighted average of the attributes (color,
bands in order to correct the estimated coefficients. In
alpha, etc.) of the four surrounding pixels is
parallel, the input image is also interpolated
computed and applied to the screen pixel. This
separately. Finally, corrected interpolated frequency
process is repeated for each pixel forming the object
sub-bands and interpolated input image are combined
being textured.
by using inverse curvelet transformation (idct) to
achieve a high resolution output image.
Bi-cubicinterpolation: In mathematics, bi-cubic
interpolation is an extension of cubic interpolation for
RESULTS:
interpolating data points on a two dimensional
regular grid. The interpolated surface is smoother
than corresponding surfaces obtained by bilinear
interpolation or nearest-neighbor interpolation. Bicubic interpolation can be accomplished using either
Lagrange polynomials, cubic splines, or cubic
convolution algorithm. In image processing, bi-cubic
interpolation
is
often
chosen
over
bilinear
interpolation or nearest neighbor in image resampling, when speed is not an issue.
PROPOSED
IMAGE
RESOLUTION
ENHANCEMENT:As the curvelet transformation
consists of four part i.e., Sub-band decomposition,
(c) Original low resolution Baboon’s image. (b)
Smooth
Ridgelet
Bicubic interpolated image. (c) Super resolved image
transformation. Mostly useful to find the edges of the
using WZP. (d) Proposed technique. (e) Curvelet
images, where every frequency component is being
With Zeropad
partitioning,
Normalization,
tested under this process. For multi resolution
analysis it is advanced one compared to Wavelet
ISSN: 2231-5381
http://www.ijettjournal.org
Page 18
International Journal of Engineering Trends and Technology (IJETT) – Volume 15 Number 1 – Sep 2014
Comparison analysis:
PSNR
TECHNIQUES/IMAGES
LENA
BABOON
BARBARA
BILINEAR
26.34
20.51
25.38
BICUBIC
26.86
20.61
28.93
WAVELET
28.84
21.47
30.44
34.82
33.06
33.06
40.11
38.1
39.26
ZEROPADDING
DWT&SWT
DECOMPOSITION
PROPOSED
TECHNIQUE
CONCLUSION:
This work proposed an image resolution
enhancement technique based on the interpolation of
the high frequency sub bands obtained by Discrete
Curvelet
Transformation,
correcting
the
high
frequency sub band estimation by FDCT frequency
sub bands, and the input image. The proposed
technique uses FDCT to decompose an image into
different sub bands, and then the high frequency sub
band images have been interpolated. Afterwards all
these images have been combined using IFDCT to
generate a super resolved imaged. The proposed
technique has been tested on well-known benchmark
[3] E. Provenzi, C. Gatta, M. Fierro, and A. Rizzi, “A
spatially variant white- patch and gray-world method
for color image enhancement driven bylocal
contrast,” vol. 30, no. 10, pp. 1757–1770, 2008.
[4] Ø. Kolås, I. Farup, and A. Rizzi, “Spatio-temporal
retinex-inspired envelope with stochastic sampling: A
framework for spatial color algorithms,” J. Imag. Sci.
Technol., vol. 55, no. 4, pp. 1–10, 2011.
[5] H. A. Chipman, E. D. Kolaczyk, and R. E.
McCulloch, “Adaptive bayesian wavelet shrinkage,”
J. Amer. Stat. Assoc., vol. 92, no. 440, pp. 1413–
1421, 1997.
[6] A. Chambolle, R. De Vore, N.-Y. Lee, and B.
Lucier, “Nonlinear wavelet image Processing:
Variational problems, compression, and noise
removal through wavelet shrinkage,” IEEE Trans.
Image Process., vol. 7, no. 3,pp. 319–335, Mar. 1998.
[7] D. Cho, T. D. Bui, and G. Chen, “Image
denoising based on wavelet shrinkage using neighbor
and level dependency,” Int. J. Wavelets,
Multiresolution Inf. Process., vol. 7, no. 3, pp. 299–
311, May 2009.
[8] E. P. Simoncelli and W. T. Freeman, “The
steerable pyramid: A flexible architecture for multiscale derivative computation,” in Proc. 2nd Annu.
Int. Conf. Image Process., Oct. 1995, pp. 444–447.
[9] J. E. Fowler, “The redundant discrete wavelet
transform and additive noise,” Mississippi State
ERC, Mississippi State University, Tech. Rep.
MSSU-COE-ERC-04-04, Mar. 2004.
images, where their PSNR and visual results show
the superiority of proposed technique over the
conventional
and
state-of-art
image
resolution
enhancement techniques.
REFERENCES:
[1] M. Fierro, W.-J. Kyung, and Y.-H. Ha, “Dual-tree
complex wavelet transform based denoising for
random spray image enahcement meth- ods,” in Proc.
6th Eur. Conf. Colour Graph., Imag. Vis., 2012, pp.
194–199.
[2] E. Provenzi, M. Fierro, A. Rizzi, L. De Carli, D.
Gadia, and D. Marini, “Random spray retinex: A new
retinex implementation to investigate the local
properties of the model,” vol. 16, no. 1, pp. 162–171,
Jan. 2007.
ISSN: 2231-5381
[10] X. Li and M. T. Orchard, “New edge-directed
interpolation,” IEEE Trans. Image Process., vol. 10,
no. 10, pp. 1521–1527, Oct. 2001.
[11] K. Kinebuchi, D. D. Muresan, and R. G.
Baraniuk, “Wavelet- based statistical signal
processing using hidden Markov models,” in Proc.
Int. Conf. Acoust., Speech, Signal Process., 2001,
vol. 3, pp. 7–11.
[12] S. Zhao, H. Han, and S. Peng, “Wavelet domain
HMT-based image super resolution,” in Proc. IEEE
Int. Conf. Image Process., Sep. 2003, vol. 2, pp. 933–
936.
[13] A. Temizel and T. Vlachos, “Wavelet domain
image resolution en- hancement using cyclespinning,” Electron. Lett., vol. 41, no. 3, pp. 119–
121, Feb. 3, 2005.
http://www.ijettjournal.org
Page 19
International Journal of Engineering Trends and Technology (IJETT) – Volume 15 Number 1 – Sep 2014
[14] A. Temizel and T. Vlachos, “Image resolution
upscaling in the wavelet domain using directional
cycle spinning,” J. Electron. Imag., vol. 14, no. 4,
2005.
CH.PULLARAO working as
faculty in LBRCE and has a complete experience of 4
years. He completed his masters from NIT Warangal.
His areas of interests are Signal Processing and
Image processing. His publications are 1.)
Significance of Textural features in Aerial Images.
2.) Target Recognition Using Multi Resolution
Analysis in Camouflaged Images. 3.) Analysis
ofAerial Images.
T.Prasad has completed his B.Tech. in ECE at
Sarojini Institute of Technology, Telaprolu. At
present he is studing M.Tech in Systems and Signal
Processing(SSP) at Lakireddy Bali Reddy College of
Engineering. Her areas of interest include image
processing, Signal processing.
ISSN: 2231-5381
http://www.ijettjournal.org
Page 20
Download