Upload
international-journal-of-technology-and-computing-ijtc
View
223
Download
0
Embed Size (px)
Citation preview
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
1/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
9
ADAPTIVE GAUSSIAN FILTER BASED IMAGE
RECOVERY USING LOCAL SEGMENTATION
SumeetWalia1, Sachin Majithia2
1Department of Information Technology Chandigarh Engineering College, Landran Mohali, Punjab, India
E-mail: [email protected]
2,Department of Information Technology Chandigarh Engineering College, Landran Mohali, Punjab, India
E-mail: [email protected]
Abstract
Image restoration is an important branch of Image processing, dealing with the reconstruction of Images by
removing noise and blur from degraded Images and making them suitable for human perception. Any Image
acquired by a device is susceptible of being degraded by the environment of acquisition and transmission.
Removal of noises from the Images is a critical issue in the field of digital Image processing. So, we propose a
new model of Image content restoration based on the Hybrid Regression which uses iterative block based
model of genetic learning by using the fittest neighbor modeling of the Image frame data under observation
using patch order kernel filtering with frame cross reference sing for dependent estimation of pattern to be
restored and determine the intensity of the filter. The paper shows evaluation of the proposed method in
comparison to previous approaches.
Keywords : Restoration, DWT, filtering, histogram filter, degradation, HVS
I. Introduction
The restoration of image as well as data forms an imperative field which is related to the group
of Image Handling. This field is basically meant to extract top notch of an image from an image
of low quality. The low quality image may be uproarious or may contain some haziness. The
processes which have been developed for the calculation of image handling are meant to
overcome a number of issues such as the restoration of an image, the division of an image, the up
gradation of an image etc. A process of information procurement is carried out for debasing the
images. It has been seen that the advantages associated with the processes undergone for the
enhancement of quality of image far exceed the implications related to the expenses or the
unpredictable nature of the rebuilding calculations. The basic motivation behind the use of the
Image restoration process is to make up for those imperfections which are debasing the image.
The corruption present in the image can be in any form such as in the form of commotion or the
misfocus of the camera. When the corruption is due to the movement of obscure, it is easy to
think of some idea to gauge the first image by fixing the corrupted image [2].
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
2/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
10
Another reason for the corruption of an image is due to commotion and the way in which such an
image can be gauged is by making adjustments in the degradation which is caused because of
commotion. In the present work, few systems that are included as a part of the field of Image
handling for the process of restoring an image are discussed. In the present time, when digital
cameras which have good focussing level and the other innovations in the field of digital media,there exists a number of cases where the image or video will still be debased because of
corruption. For carrying out the reconstruction of such an image, image rebuilding can be
considered as an emergency approach.
The hypothesis clarifies that in a case in which the image is undermined; it becomes a key issue
to make some sort of move in order to repair that image. The correspondence is thought of as a
Fundamental intend for the restoration of an image in case of an emergency. With the help of
correspondence, an element grants permission to the other elements so that they can comprehend
an image and the impacts present in the image by utilising an arrangement of the experiences of
the individuals who are prompting their transactions from the association or the company as in[4].
A. Image Processing
The process of image rebuilding was first adopted in 1950’s. Some of the areas in which image
rebuilding finds application are: investigation of experiments [9], examinations related to law,
film making. Another area of application of image rebuilding is the shopper photography as well
as feature unravelling. However, the primary zone in which it is utilised is the recreation of
images in radio space science or in radar imaging. The image rebuilding process [9] makes use
of the earlier information related to the degradation. It also takes into account the corruption in
the images and then the converse procedure is applied to them. It takes into consideration thegoodness of the image as the target criteria. The contortion is then demonstrated in the form of
clamour or in the form of capacity of degradation. With the aim of restoration of an image from
the commotion model, a variety of channels including the middle channel as well as homo
morphic channel are employed. In order to get rid of the intermittent commotions, the various
channels which are utilised are the middle channel, the Butterworth low pass channel or band
reject channels. For the purpose of restoration of the image in case of a direct degradation, the
processes of reverse as well as pseudo backwards shifting or the weiner separation are employed.
B. Image Restoration
It is concerned with the reconstruction and estimation of the uncorrupted image from blurred and
noisy images. It tries to perform an operation on the image that is the inverse of the
imperfections in the image formation system. In the use of image restoration methods, the
characteristics of the degrading system and the noise are assumed to be known unknown. Image
restoration algorithms distinguish themselves from image enhancement methods in that they are
based on models for the degrading process and for the ideal images. For those cases where a
fairly accurate blur model is available, powerful restoration algorithms can be arrived at.
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
3/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
11
Unfortunately, in numerous practical cases of interest the modeling of the blur is unfeasible,
rendering restoration impossible.
Fig. 1: Processing of Image Restoration
Figure 1 represents the Technique for processing of image restoration. Image Processing is the
processing of Images and Videos using mathematical operations where input is an image, such as
photograph or a Video Frame, output of Image Processing can be a Image or some set of
characters. Most Image Processing techniques involve treating the Image as a two-dimensional
signal and applying standard signal processing technique to it. The restoration systems use
different sorts of channels for accomplishing best execution, similar to converse channel, wiener
channel, obliged minimum square channel, Histogram Adaptive Fuzzy channel, Min-max
Detector Based channel and Centre Weighted Mean channel and so for others [14].
The limited validity of blur models is often a factor of disappointment, but one should realize
that if none of the blur models are applicable, the corrupted image may well be beyond
restoration. Therefore, no matter how powerful blur identification and restoration algorithms are,
the objective when capturing an image undeniably is to avoid the need for restoring the image.
Image restoration deals with images recorded in the presence of one or more sources of
degradation. Many sources of degradation are present in imaging systems [12].Some types affect
only the gray levels at the individual picture points without introducing spatial blur. They are
called point degradation. Other types which do involve blur are spatial degradations. Blurring is
a form of bandwidth reduction of an ideal image caused by imperfect image formation process.
II. Literature Review
In order to enhance the nature of the pixonal Images, [11] has proposed two augmentations for
the images based on pixon. The first augmentation is achieved with the help of
bicubicinterpolation and the next augmentation is based on expansions which are achieved by
making use of the technique of fluffy shifting. With the aim of removal of noise from the pixonal
Degraded Image
Knowledge of
Image Creation
Develop InverseDe radation
Apply Inverse
De radation
DevelopDegradatio
Input Output
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
4/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
12
image, there is the provision for connecting the differential mathematical functions on it. In an
attempt to reduce the various noises and to restore the image, the systems of comparison
employing fluffy separation and incomplete differential systems are consolidated. The results
obtained depict that the proposed scheme is better than the current systems which are based on
the PDE strategy or wavelet systems.
In [21], the authors have suggested a model that can be used for carrying out mapping between
the spaces by employing the coupled lexicon learning. The technique makes use of the references
which have been learnt for the purpose of recuperating the patch in case of the patches having
low determination. In the work proposed, two bearings have been utilised for the enhancement of
productivity and in order to accelerate the calculations:
1) The LR Image patches are selectively transformed by considering the characteristic
insights of an image.
2)
Learning of a neural network that will result in quick representation derivation.
The image patches can undergo processing adaptively by taking into account the standard
deviation of threshold by making use of the neural networks. In [25], the authors have suggested
a strategy for the preparation of coupled word reference for the case of super resolution of single
image. With the help of some inadequate representations, the proposed strategy can relate the
patches of low and high determination. The preparing power of the coupled lexicon which is
achieved from the patch of LR can result in the reproduction of HR patch of the word reference.
The calculation so performed leads to the enhancement of the exactness of recuperation. Apart
from this, in this time the relics of the recuperation can also be uprooted.
A Bayesian super determination calculation is proposed in this paper with characteristic Image
measurements by utilizing generative plans for high determination Image rebuilding Image
estimation through testing. It uses the characteristic Image measurements for Image SR with
utilizing adaptable high request Markov Random Field model. Field-of-master model is utilized
to take in the former model from normal Images. The creators have proposed a completely
Bayesian approach, that partners a former learning on concealed high determination Image and
in addition the noise level into the structure in regular ways in. High request MRF model is
utilized for catching the high determination Image insights without utilizing priors and for
displaying the normal Image measurements additionally empowers generative inspecting. The
Bayesian least mean square mistake (MMSE) criterion is used to shape estimate of HR Image
[25]. This MMSE method does not oblige impromptu change for accomplishing alluring
rebuilding execution. MMSE criteria are less touchy for the nearby minima in the arrangement
space than the MAP. Trial tests demonstrate that the proposed system can create preferable
results over the cutting edge SR calculations. To begin with restoration the model was analysed
for individual storm streak and snow. This model is then fit to a component and is used to
recognize rain or snow streaks first in repeat space, and the area result is then traded to Image
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
5/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
13
space. The shortcoming is that it is not material for light rain, ensuing to the sample moulded in
repeat space is not specific.
The Phase technique uses old techniques that are available in video outlines with noise.
According to the phases of video chains distinctive sorts of curios are distinguished which may
contain obscure delivered amid obtaining, post-preparing an interpreting. For rebuilding reason,
every one of these sorts are connected on edge to watch the impacts of these sorts on casing.
Division calculation is utilized which is in view of visual consideration model for recognizing
piece of haziness on entire edge and the deliberate obscure of the foundation. Trial test
demonstrates that it can illuminate diverse sorts of antiquities furthermore the correspondence
between metric qualities and judgements of human perception. Viability of these measurements
are tried chiefly against subjective judgements review of techniques for digital restoration of
images. Some examples of restoration were included to illustrate the methods discussed in [31].
A special phase transformation to extract linear features successfully from satellite images was
made. Many motion-blurred image restoration methods were proposed recent years, and themethod adopted a multiphase spatial and spectral approach for restoration.
In [32] this method the creators have proposed a non-neighbourhood portion relapse model for
Image and video restoration. This model joins the non-nearby self-comparability and
neighbourhood auxiliary normality properties for solid and vigorous estimation of normal
Images. The proposed technique performs Image and videodeblurring, denoising and super-
determination chip away at Images. Neighbourhood basic normality is utilized for watching the
patches of Images that has customary structures where accurate judgment of pixel values through
relapse is conceivable. The non-nearby self-closeness is primarily in light of the perception of
Image fix that can rehash themselves in Images and videos. A similar approach based on Group-Based Sparse Representation (GSR) system was proposed. As opposed to utilizing fix as an open
unit they misused the idea of gathering as an essential unit of inadequate representation which is
a mix of nonlocal patches which has comparable structure. The proposed GSR displaying
contains three folds [33]. The principal fold speak to the normal Images scantily in space of
gathering which portrays the non-nearby self-similitude and characteristic neighborhood sparsity
of Images all the while in a brought together structure. In second crease, instead of utilizing
lexicon gaining from normal Images, a versatile gathering word reference learning system is
composed with low intricacy.
III. Problem Formulation
In the review of various techniques, the problem of applying recovery to large set of images is
the high computing time required for processing and the requirement for a large matrix data. In
this research work, we propose a fast image recovery algorithm by dividing the image into block
of pixels and applying to each block instead of the entire image. The previous approach had
problems relating to the recovery of data with artifact generation, this occurred primarily due to
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
6/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
14
the non-regularized filtering and global filtering. This resulted in a lower PSNR, due to the fact
that the saturation of data with low MSE, was filtered resulting in disorientation of pixel data.
IV. Proposed Work Methodology
[1]
Select the image for recovery from database and pre process or dimensioning for analysis[2] Calculate Gaussian after calculation of Gaussian and deviation values, the kernel filter is
designed using non local gradients according to Gaussian calculations.
[3] Gaussian based deviation using TV estimation
[4] The kernel filter derived from above system is convoluted with image and the intensity of the
filter is set according to the change in the block based standard deviation.
[5] Calculate Gaussian for the image at final stage
[6] Filter the image using iterative kernel with average power ratio Based Segment Enhancement
[7] On basis of avg power ratio segmentation should increase or decrease. Avg power ratio will
change for particular block, where the proportion is constant that is maximum. Power ratio is
segmented of each block. Then maximum limit is calculated.
[8] Continue recovery till the power ratio is stable with low deviation for all the blocks under
observation
[9] Calculate parameter and analyze the results
V. Quality Metrics Used
A. Mean-squared error (MSE)
Mean-squared error (MSE) has been the dominant quantitative performancemetric in the field of
signal processing. It remains the standard criterion for the assessment of signal quality and
fidelity; it is the method of choice for comparing competing signal processing methods
andsystems, and, perhaps most importantly, it is the nearly ubiquitous preference of design
engineers seeking to optimize signal processing algorithms.
1 1
2
0 0
( , ) '( , )W H
x y
f x y f x y
MSE WH
…………………(5.1)
B. Peak Signal to Noise Ratio (PSNR)
Peak signal-to-noise proportion, frequently condensed PSNR, is a designing term for the
proportion between the greatest conceivable force of a sign and the force of tainting commotion
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
7/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
15
that influences the devotion of its representation. Since numerous signs have a wide element
range, PSNR is typically communicated as far as the logarithmic decibel scale.
10
25520log PSNR
MSE
………………………………(5.2)
C. Block Diagram
Fig. 2: shows the proposed system flow diagram
VI. Results and Discussion
The below section shows the comparison of the proposed method with that of the previous
systems using visual output comparison and metric based evaluation of PSNR, MSE and Entropy
values.
Select the image
Analyze for image quality and pres
process
Calculate Gaussian and performfiltering using TV based Kernel
Avg. power ratio based segmentation
enhancement
Continue segment enhancement with kernel filtering till low
deviation is achieved
Get the total recovered image
Get average
power deviation
Perform evaluation of the system in
all environments
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
8/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
16
Table 1: Tabular representation of PSNR Values at Different Gaussian Noise Level
Guassian
Noise Levels
Median
Filter
Avg. Filter Hist.
Eq.Filter
Weiner Filter Prop. Filter
PSNR PSNR PSNR PSNR PSNR
5 8.487671 8.517238 13.88805 10.57732 15.00604
10 8.70839 7.59384 10.794 9.17899 13.38049
15 7.654813 6.670538 9.972098 10.642486 12.676373
20 6.956329 5.815018 9.066818 8.911423 11.896088
25 6.242164 5.187538 8.017599 8.183708 11.147752
Fig. 4: Graphical representation of PSNR Values at Different Gaussian Noise Levels
0
5
10
15
20
25
5 10 15 20 25
P S N R
( I n d B )
Gaussian Noise Levels
Median Filter
Average Filter
Hist. Eq. Filter
Weiner Filter
Prpposed Filter
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
9/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
17
Table 2: Tabular representation of S.D Values at Different Gaussian Noise Levels
GuassianNoise Levels
MedianFilter
Avg. Filter Hist.
Eq.Filter
Weiner Filter Prop. Filter
Std. Dev Std. Dev Std. Dev Std. Dev Std. Dev
5 13.487671 13.517238 18.88805 15.57732 20.00604
10 13.70839 12.59384 15.794 14.17899 18.38049
15 12.654813 11.670538 14.972098 15.642486 17.676373
20 11.95633 10.81502 14.06682 13.91142 16.89609
2511.24216 10.18754
13.0176
13.18371
16.14775
Fig. 6: Graphical representation of S.D Values at Different Gaussian Noise Levels
0
5
10
15
20
25
5 10 15 20 25
S t a n d a r d D e v i a t i o n
Gaussian Noise Levels
Median Filter
Average Filter
Hist. Eq. Filter
Weiner Filter
Proposed Filter
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
10/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
18
Table 3: Tabular representation of MSE Values at Different Gaussian Noise Levels
Guassian
Noise Levels Median Filter Avg. Filter Hist.
Eq.Filter
Weiner Filter Prop. Filter
MSE MSE MSE MSE MSE
5 16.037671 16.067238 21.43805 18.12732 22.55604
10 16.25839 15.14384 18.344 16.72899 20.93049
15 15.204813 14.220538 17.522098 18.192486 20.226373
20 14.50633 13.36502 16.61682 16.46142 19.44609
25 13.79216 12.73754 15.5676 15.73371 18.69775
Fig. 8: Graphical representation of MSE Values at Different Gaussian Noise Levels
VII. Conclusion and Future Scope
Image restoration is an important branch of image processing, dealing with the reconstruction of
images by removing noise and blur from degraded images and making them suitable for human
perception. Any image acquired by a device is susceptible of being degraded by the environment
0
5
10
15
20
25
5 10 15 20 25
M S E
Gaussian Noise Levels
Median Filter
Average Filter
Hist. Eq. Filter
Weiner Filter
Proposed Filter
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
11/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
19
of acquisition and transmission. Therefore, a fundamental problem in the image processing is the
improvement of their quality through the reduction of the noise that they can contain being often
known as "cleaning of images". The goal of the restoration approach is to improve the given
image so that it is suitable for further processing. Removal of noises from the images is a critical
issue in the field of digital image processing. Various filters and techniques are used in imagerestoration to restore the corrupted image to its original form. The restoration results in the
improved quality of image. The types of noises are explained and discussed along with their
probability density functions (PDF).Various spatial filtering techniques are used for reducing
these noises from images. The proposed system work to perfect the images and make them free
from all the given disturbances, with truth evaluation using various quality metrics showing the
high performance capacity of the proposed system above the previous surveyed systems.
In future the proposed work can be implemented in real-time operating system using VLSI
design in the microchip for onboard use in mobiles and portable laptop devices, the method can
also be modified to use in devices with limited RAM space.
REFERENCES
[1] Douglas, S.C.“Introduction to Adaptive Filters”, Digital Signal Processing Handbook
[2] Jian Zhang, Debin Zhao, Wen Gao, (August 2014) ” Group-Based Sparse Representation for Image
Restoration”, IEEE Transactions On Image Processing, Vol. 23, No. 8.
[3] BiswaRanjanMohapatra, Ansuman Mishra, Sarat Kumar Rout, (March 2014 )“A Comprehensive Review onImage Restoration Techniques”, International Journal of Research in Advent Technology, Vol.2, No.3, E-
ISSN: 2321-9637
[4] Baochen Jiang, Aiping Yang, Chengyou Wang and ZhengxinHou,(2014) “Comparison of Motion-blurred
Image Restoration Using Wiener Filter and Spatial Difference Technique” International Journal of SignalProcessing, Image Processing and Pattern Recognition Vol.7, No.2, pp.11-22
http://dx.doi.org/10.14257/ijsip.2014.7.2.02[5] R. E. KALMAN,, (2009).“A New Approach to Linear Filtering and Prediction Problems”, International
Journal of Recent Trends in Engineering, vol 4, no. 6.
[6] Francesca Dardi ,LeonardoAbate , JeroenStessen , GiovanniRamponi , (2013) “ Causes and subjectiveevaluation of blurriness in video frames “, Elsevier Signal Processing: Image Communication 28 209 – 221.
[7] Haichao Zhang, Jianchao Yang, Yanning Zhang , Thomas S. Huang,(June 2013) “ Image and VideoRestorations via Nonlocal Kernel Regression”, IEEE Transactions On Cybernetics, Vol. 43, No. 3,
[8] Dr. A. Sri Krishna, G.SrinivasaRao and M. Sravya, (2013)“Contrast Enhancement Techniques Using
Histogram Equalization Methods on Color Images with Poor Lightning”, Vol. 11, No. 7.
[9] SatyadhyanChickerur, Aswatha Kumar,(2009) November “A Biologically Inspired Filter For Image
Restoration”, International Journal of Recent Trends in Engineering, vol 2, no. 2.[10] R. Srivastava, H. Parthasarthy, J. R. P. Gupta, and D. R. Choudhary,, (2009) “Image restoration from motion
blurred image using PDEs formalism”, Proceedings of the IEEE International Advance Computing
Conference, Patiala, India, March 6-7, pp. 61-64.
[11]
EhsanNadernejada, Mohsen Nikpour,(30 May 2012)” Image denoising using new pixon representation basedon fuzzy filtering and partial differential equations, “Elsevie, Digital Signal Processing 22 913 – 922.
[12] Pinton, Gianmarco F., Gregg E. Trahey, and Jeremy J. Dahl.6 (2011) "Sources of image degradation in
fundamental and harmonic ultrasound imaging: a nonlinear, full-wave, simulation study." IEEE transactionson ultrasonics, ferroelectrics, and frequency control 58:1272.
[13] Cai, Jian-Feng, et al. (2012)"Image restoration: Total variation, wavelet frames, and beyond." Journal of the
American Mathematical Society 25.4: 1033-1089.
[14] Varghese, Jobin, et al.(2014) "Efficient adaptive fuzzy-based switching weighted average filter for the
restoration of impulse corrupted digital images." Image Processing, IET 8.4: 199-206.
http://dx.doi.org/10.14257/ijsip.2014.7.2.02http://dx.doi.org/10.14257/ijsip.2014.7.2.02
8/17/2019 IJTC201601002-Adaptive Gaussian Filter Based Image Recovery Using Local Segmentation
12/12
INTERNATIONAL JOURNAL OF TECHNOLOGY AND COMPUTING (IJTC) ISSN-2455-099X,
Volume 2, Issue 1 January 2016.
20
[15] Esakkirajan, S., et al.(2011) "Removal of high density salt and pepper noise through modified decision based
unsymmetric trimmed median filter." Signal Processing Letters, IEEE 18.5: 287-290.
[16] Khare, Charu, and Kapil Kumar Nagwanshi.(June 2011) "Implementation and analysis of image restoration
techniques." International Journal of Computer Trends and Technology-May to (2011).
[17] Shukla, Awanish Kumar, et al. "An improved directional weighted median filter for restoration of images
corrupted with high density impulse noise(2014) "Optimization, Reliabilty, and Information Technology
(ICROIT), 2014 International Conference on. IEEE.[18] W. Quan and W. N. Zhang, (2011) “Restoration of motion- blurred star image based on Wiener filter”,
Proceedings of the 4th International Conference on Intelligent Computation Technology and Automation,
Shenzhen, China, March 28-29, pp. 691-694.[19] X. Zhang, H. Li, Y. Qi, W.K. Leow and T.K. Ng, , (2006)"Rain Removal in Video by Combining Temporal
and Chromatic Properties," in IEEE Int. Conf. Multimedia and Expo, pp. 461-464.
[20] BalvantSingh , Ravi Shankar Mishra, PuranGour (2009)Analysis of Contrast Enhancement Techniques ForUnderwater Image, vol. 1, pp. 328-335.
[21] Jianchao Yang, Zhaowen Wang, Zhe Lin, Scott Cohen, Thomas Huang, “Coupled Dictionary Training for
Image Super-Resolution”, IEEE Transactions On Image Processing, Vol. 21, No. 8, August 2012.
[22] KshitizGarg and Shree K. Nayar, “Vision and Rain”, International Journal of Computer Vision 75(1), 3– 27,
February 2007[23] Reddy, P. HarshaVardhan, B. Karthik, and TVU Kiran Kumar. "Restoration of Image Using DCT and
Modified Decision Based Unsymmetrical Trimmed Median Filter."
International Journal of Research
2.4(2015): 846-849.
[24] Rainer Poisel and Simon Tjoa,”A Comprehensive Literature Review of File Carving”, 2013 International
Conference on Availability, Reliability and Security.[25] Haichao Zhang, Yanning Zhang , Haisen Li, Thomas S. Huang,” Generative Bayesian Image Super
Resolution with natural image prior ”, IEEE Transactions On Image Processing, Vol. 21, No. 9, September
2012.[26] Taeg Sang Cho, C. Lawrence Zitnick, Neel Joshi, Sing Bing Kang, Richard Szeliski, and William T. Freeman
,” Image Restoration by Matching Gradient Distributions”, IEEE Transactions On Pattern Analysis And
Machine Intelligence, Vol. 34, No. 4, April 2012.
[27] Ming Zhou, Zhichao Zhu, Rong, Deng, ShuaiFang,Rain, “Detection and Removal of Sequential Images”,
IEEE Chinese Control and Decision Conference (CCDC), pp. 615-618, 2011.[28] Peter C. Barnum ,SrinivasaNarasimhan ,Takeo Kanade “Analysis of Rain and Snow in Frequency Space”,
Springer International Journal of Computer Vision 86,256-274, 2010.[29]
R. E. KALMAN, “A New Approach to Linear Filtering and Prediction Problems”, International Journal of
Recent Trends in Engineering, vol 4, no. 6, (2009).[30] K. Garg and S.K. Nayar, "Detection and removal of rain from videos," in IEEE Conf. Computer Vision and
Pattern Recognition, vol. 1, pp. 528-535, 2004.
[31] K. Jain, "Fundamentals of Digital Image Processing", Engelwood Cliff, N. J.: Print ice Hall, (2006).
[32] Guillemot, Christine, and Olivier Le Meur. "Image inpainting: Overview and recent advances." Signal
Processing Magazine, IEEE 31.1 (2014): 127-144.
[33] NasirMemon and Anandabrata Pal” Automated Reassembly of File Fragmented Images Using GreedyAlgorithms”, IEEE Transactions of image processing, VOL. 15, NO. 2, FEBRUARY 2006.
[34] H. Taub, D.L. Schilling,(1991) "Principles of Communication Systems", TMH.
[35] Piotr S. Windyaga,(2001) January) "Fast Implusive Noise Removal", IEEE Trans. On Image Processing,
vol.1 0, no, 1.