SlideShare une entreprise Scribd logo
1  sur  6
Télécharger pour lire hors ligne
Full Paper
ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013

Reduced Ordering Based Approach to Impulsive
Noise Suppression in Color Images
Bogdan Smolka, Krystyna Malik
Silesian University of Technology, Akademicka 16, 44-100 Gliwice, Poland
{Bogdan.Smolka; Krystyna.Malik}@polsl.pl
Abstract—In this paper a novel filtering design intended for
the impulsive noise removal in color images is presented.
The described scheme utilizes the rank weighted cumulated
distances between the pixels belonging to the local filtering
window. The impulse detection scheme is based on the
difference between the aggregated weighted distances assigned
to the central pixel of the window and the minimum value,
which corresponds to the rank weighted vector median. If the
difference exceeds an adaptively determined threshold value,
then the processed pixel is replaced by the mean of the
neighboring pixels, which were found to be not corrupted,
otherwise it is retained. The important feature of the described
filtering framework is its ability to effectively suppress
impulsive noise, while preserving fine image details. The
comparison with the state-of-the-art denoising schemes
revealed that the proposed filter yields better restoration
results in terms of objective restoration quality measures.

noise removal in color images is based on the order statistics
[7]. These filters perform the vector ordering of the set of
pixels from the filtering window to determine the filtering
output. The widely used reduced vector ordering is based
on assigning a dissimilarity measure to each color pixel from
the filtering window [1,8,9]. The aggregated dissimilarity
measure assigned to pixel xi is defined as
(1)
where
denotes the distance between pixels in a given
color space. The values of Ri, (i = 1,...,n) are then sorted and
the vectors x1, x2, ..., xn are correspondingly ordered
(2)
where < denotes the order relation between vectors and
R(k) denotes the k-th smallest value of R.
Many denoising techniques define the vector x(1) in (2)
as their output, since vectors that diverge significantly from
the samples of w appear in the higher indexed locations in
their ordered sequence.
One of the most widely used noise reduction techniques
is the Vector Median Filter (VMF), whose output is the vector
x(1) from W, for which the sum of distances to all other vectors
is minimized
(3)

Index Terms—color image processing, noise reduction, image
enhancement

I. INTRODUCTION
Quite often the quality of color images is degraded by
various types of noise, whose removal is required to enable
the success of further steps in the image processing pipeline.
Noise, arising from a variety of sources, is inherent to
electronic image sensors and therefore the noisy signal has
to be processed by a suitable filtering algorithm that reduces
the noise component, while preserving original image details
[1-3].
In this work we focus on a special kind of distortion,
which is introduced by impulsive noise caused by
malfunctioning sensors, faulty memory locations in hardware,
aging of the storage material or transmission errors due to
natural or man-made processes [4-6].
The color image will be treated as a two-dimensional matrix consisting of pixels
, where the
index u = 1,...,N indicates the pixel position on the image
domain and v denotes the number of channels. In the case
of standard color images v=3. The vector components xuk,
for k = 1,2,3 represent the RGB color channels values.
Generally, filtering operators work on the assumption that
the local image features can be extracted from a small image
region centered at pixel xu, called a sliding filtering window
and denoted as w. Thus, the output of the filtering operation
will depend only on the n samples contained within the
filtering window centered at xu, which will be also denoted
for convenience as , so that .
The majority of the nonlinear filters used for impulsive
© 2013 ACEEE
DOI: 01.IJSIP.4.3. 1255

where
denotes Euclidean norm. The VMF is the most
popular operator intended for smoothing out the impulses
injected into the color image by the noise process. This filter
is very efficient at reducing the impulses, preserves sharp
edges and linear trends. However, the drawback of the VMF
and other filters based on vector ordering lies in introducing
excessive smoothing, which results in an extensive blurring
of the output images. Thus, the filters based on vector ordering do not preserve fine image structures, which are treated
as noise and therefore generally show a tendency to the
blurring of image details and generation of color artifacts.
In order to alleviate the problem of image smoothing
various switching filters, that replace only the corrupted
pixels have been proposed [10-13]. The efficiency of a
switching filter depends both on the quality of the impulse
detection scheme and on the applied restoration framework,
which replaces the detected impulses with estimates derived
from the samples belonging to a local processing window.
In this paper a novel switching filter is proposed. The
main advantage of the proposed approach is its ability to
suppress the noise component, while preserving fine image
44
Full Paper
ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013
details. The structure of the filter is based on the reduced
ordering statistics and is characterized by a low
computational complexity, which enables the adoption of the
new technique in real-time applications.

proposed method. It can be observed that the map of the
detected noise correlates very well with the real contamination measured as the Euclidean distance between the original and corrupted pixels in the RGB color space normalized to
the range < 0,255>. Figure 3 exhibits the correlation between the real and the detected impulsiveness of the image
pixels using the test images Goldhill and Parrots. The results reveal a high correlation between the real and detected
noise.

II. RANK WEIGHTED FILTERING DESIGN
The reduced ordering schemes are based on the sum of
the dissimilarity measures between a given pixel and the
samples from the local filtering window w. In this way, the
output of the vector median filter is the pixel whose average
distance to other pixels is minimized.
The distances
between the pixel xiand all
other pixels w belonging to can be arranged into a sequence
(4)
and the ranks of the ordered distances can be used for the
evaluation of the aggregated distances in (1).
If stands for the rank of a given distance, then
will
denote the corresponding distance value and instead of the
aggregated distances in (1) a weighted sum of distances,
utilizing the distance ranks, can be composed
(5)

Figure 1. Illustration of the noise detection efficiency using the
Parrots image corrupted by impulsive noise with p=0.1 intensity.

where f(r) is a decreasing weighting function of the distance
rank r. Then, the rank weighted sums of distances Di can be
sorted and a new sequence of vectors is obtained
(6)
where the vector
is the output of the Rank Weighted
Vector Median Filter (RWVMF). Applying a step-like function
(7)
the Sharpening Vector Median Filter (SVMF) presented in
[14] is obtained.
Extensive experiments revealed that satisfactory denoising
results are achieved using monotonously decreasing function
f(r) = 1/r [15-17]. The weighting function decreases the
influence of large distances introduced by the outliers injected
by the noise process, which enables to efficiently remove
the impulsive noise while enhancing the image edges.

Figure 2. Illustration of the noise detection efficiency using the
Goldhill image corrupted by noise with p=0.1 intensity.

Figure 4 shows the histograms of the corresponding values
of D1, D(1) and the measure of impulsiveness D1 - D(1) for
different levels of noise intensity using the Parrots test image.
As can be seen, with increasing level of noise intensity, the
histograms are shifted towards higher values, which is
confirmed by the mean value of D1 - D(1) marked with a red
line on the plots.
The structure of the proposed switching filter is quite
simple. If the difference D1 - D(1), which measures the pixel
corruption, exceeds a given threshold value , then a pixel is
declared as corrupted by a noise process, otherwise it is
treated as not disturbed

III. ADAPTIVE SWITCHING FILTER
In order to decide whether a pixel of a color image is corrupted by impulsive noise, the difference between the cumulated weighted distance D1 assigned to the central pixel of
the filtering window and the value of D(1) corresponding to
the rank weighted vector median filter output can be used
[15-17] and the strength of the impulsive contamination can
be estimated as the difference between D1 and D(1).
Figures 1 and 2 show examples of the detected noise using parts of the color test images Parrots and Goldhill corrupted by impulsive noise, whose intensity p denotes the
percentage of corrupted pixels. The visual comparison of the
real and detected noise confirm the good efficiency of the
© 2013 ACEEE
DOI: 01.IJSIP.4.3.1255

(8)
45
Full Paper
ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013
measure D1 - D(1) computed for all image pixels and denoted
as , (see also Fig. 4). The dependence between the average
impulsiveness measure
and noise intensity level p is
shown in Fig. 7.

Figure 3. Illustration of the correlation between the real noise and
the output of the proposed noise detector for different levels of
noise intensity. The correlation coefficient p is provided for each
plot.

where y1 is the switching filter output, x1 is the central pixel of
the filtering window and
is the Arithmetic Mean Filter
output computed using only the pixels found by the detector
to be not corrupted by the noise process. If all neighbors of
the central pixel x1 of the filtering window are declared as
corrupted, then the VMF applied to all pixels from w is taken
as the filter output.
Of course, the efficiency of the switching scheme is
dependent mainly on the value of the thresholding parameter.
If the threshold T is too low, the filter will be replacing
uncorrupted pixels and much of the image details will be lost.
On the other hand, if the value of T is too high, many
corrupted pixels will pass the filter without being processed.
As could be expected, the optimal setting of T depends
on the contamination intensity. As can be observed in Fig. 6,
which shows the dependence between the optimal threshold
. value and the noise contamination level p evaluated using
a set of test images shown in Fig. 5, the threshold yielding
the best PSNR value is decreasing with increasing noise
contamination level and does not depend significantly on
the image structure.
Therefore, the thresholding parameter T needs to be
adjusted to the noise intensity level. The experiments
performed using the set of images shown in Fig. 5, indicate
an approximately linear dependence between the
contamination ratio p and the mean value of the impulsiveness
© 2013 ACEEE
DOI: 01.IJSIP.4.3.1255

Figure 4. Histograms of the values of D1, D(1) and D1 - D(1) created
using the PARROTS image contaminated by the impulsive noise
with different intensities. The red line marks the mean value of a
histogram.

46
Full Paper
ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013

Figure 5. Color test images used for the construction
of an adaptive filter.

Figure 1. Dependence between the optimal threshold
and the
output of the noise detector for the images shown in Fig. 5.

IV. COMPARISON WITH E XISTING TECHNIQUES
For the evaluation of the effectiveness of the proposed
filtering technique the test images were corrupted by a
impulsive noise modeled in such a way, that the noisy pixel
is defined as:
(10)
Figure 6. Dependence of the optimal threshold
on the
contamination level p for the images shown in Fig. 5 corrupted by
impulsive noise.

where
and
,
. In
this impulsive noise model, the affected pixels have corrupted
all three channels, which take on random values from the
interval <0,255>.
For the measurement of the restoration quality, the commonly
used Root Mean Squared Error (RMSE) expressed through
the Peak Signal to Noise Ratio (PSNR) was used, as the RMSE
is a good measure of the efficiency of impulsive noise
suppression and correlates well with other commonly used
restoration quality measures [1].
The effectiveness of the proposed adaptive technique
was compared with a set of switching filters intended for the
suppression of impulsive noise in color images. The following
filters were chosen for the comparison:
Peer Group Filter (F1), [18],
 Adaptive Center Weighted VMF (F2), [19],
Adaptive Center Weighted Directional-Distance Filter (F3),
[20],
Rank-Ordered VMF (F4), [21],
Sigma Directional Distance Filter (F5), [22].
Their parameters were set according to the
recommendations provided in the appropriate references.
The four images depicted in Fig. 9 were contaminated by
the impulsive noise with intensities ranging from 0.05 to 0.3.
It is worth noticing, that these images were not included in
the set of images used to establish the dependence expressed
by Eq. (9). The results are summarized in Tab. 1.
As can be observed the proposed filtering method yields

Figure 7. Dependence of the measure of impulsiveness
on the
contamination level for color images depicted in Fig. 5.

Combining these plots, we obtain the scatter plot depicted
in Fig. 8, which reveals a roughly linear dependence between
the optimal value of the adaptive threshold and the value
of
(9)
which can be used for the adaptive setting of the appropriate threshold value enabling high efficiency of the proposed filter for low and moderate noise contamination levels.
© 2013 ACEEE
DOI: 01.IJSIP.4.3. 1255

47
Full Paper
ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013
results significantly superior to those obtained using the
state-of-the-art denoising methods.

Figure 9. Test images used for the evaluation of the restoration
results.
TABLE I. COMPARISON OF THE PSNR VALUES OBTAINED WHEN RESTORING THE
COLOR TEST IMAGES CONTAMINATED WITH THE IMPULSIVE N OISE USING THE
PROPOSED TECHNIQUE AND SOME COMPETITIVE FILTERS
Im ag e

G o ld hi ll

L en a

Pa rr ot s

Pep per s

??

NE W

F1

F2

F3

F4

F5

0. 1
0. 2
0. 3
0. 1
0. 2
0. 3
0. 1
0. 2
0. 3
0. 1
0. 2
0. 3

37 .1 4

3 6. 19
3 2. 66
2 9. 14
3 8. 56
3 4. 54
3 0. 29
3 7. 32
3 3. 98
2 9. 43
3 7. 56
3 3. 42
2 8. 58

3 5. 47
3 1. 07
2 6. 29
3 7. 87
3 1. 93
2 6. 62
3 6. 73
3 1. 37
2 6. 03
3 6. 71
3 0. 79
2 5. 10

3 3. 94
3 0. 53
2 6. 84
3 6. 34
3 1. 69
2 7. 03
3 6. 78
3 1. 83
2 6. 73
3 5. 09
3 0. 27
2 4. 89

3 5. 54
3 2. 26
2 8. 52
3 8. 12
3 3. 98
2 9. 65
3 6. 29
3 3. 09
2 8. 59
3 7. 07
3 2. 75
2 7. 80

3 4. 94
2 9. 57
2 4. 13
3 6. 07
2 9. 46
2 3. 75
3 6. 70
2 9. 82
2 3. 81
3 4. 36
2 7. 59
2 1. 98

33 .8 5
31 .3 3
39 .8 1
36 .3 5
33 .3 5
38 .6 0
35 .5 5
32 .2 1
41 .6 2
38 .4 5
34 .3 1

The results summarized in Tab. 1 are confirmed by the
subjective analysis of the results depicted in Fig. 10, which
shows the restoration quality achieved using the proposed
filter as compared with other reference filters.
IV. CONCLUSIONS
In this paper a new switching filtering design has been
proposed. The filter is based on the weighted cumulative
distances between pixels, which are used for the detection of
samples corrupted by impulsive noise process. The experiments performed on test images indicate a high efficiency of
the proposed filtering design. Moreover, the new filter has a
low computational complexity and simple structure, which
makes it attractive for real-time applications.

Figure 10. Comparison of the efficiency of the proposed technique
with other methods using a part of the color test image GOLDHILL
contaminated by impulsive noise with intensity p = 0.1

[5] J. Zheng, K. Valavanis, and J. Gauch, “Noise removal from
color images,” Journal of Intelligent and Robotic Systems,
vol. 7, no. 1, pp. 257–285, August 1993.
[6] H. Faraji and W.J. MacLean, “CCD noise removal in digital
images,” IEEE Transactions on Image Processing, vol. 15, no.
9, pp. 2676–2685, 2006.
[7] J. Astola, P. Haavisto, and Y. Neuvo, “Vector median filters,”
Proceedings of the IEEE, vol. 78, no. 4, pp. 678–689, 1990.
[8] B. Smolka, K. Plataniotis, and A. Venetsanopoulos, Nonlinear
Signal and Image Processing: Theory, Methods, and
Applications. CRC Press, 2004, chapter: Nonlinear techniques
for color image processing, pp. 445–505.
[9] B. Smolka and A. Venetsanopoulos, Color Image Processing:
Methods and Applications. CRC Press, 2006, chapter: Noise
reduction and edge detection in color images, pp. 75–100.
[10] R. Lukac, “Adaptive vector median filtering,” Pattern
Recognition Letters, vol. 24, no. 12, pp. 1889–1899, 2003.
[11] B. Smolka, A. Chydzinski, K. W. Wojciechowski, K. N.
Plataniotis, and A. N. Venetsanopoulos, “On the reduction of
impulsive noise in multichannel image processing,” Optical
Engineering, vol. 40, no. 6, pp. 902–908, 2001.
[12] B. Smolka, K.N. Plataniotis, A. Chydzinski, M. Szczepanski,

ACKNOWLEDGMENT
This work was supported by the Polish National Science
Center (NCN) under the Grant: DEC-2012/05/B/ST6/03428.
REFERENCES
[1] K. Plataniotis and A. Venetsanopoulos, Color Image Processing
and Applications. Springer Verlag, 2000.
[2] R. Lukac, B. Smolka, K. Martin, K. Plataniotis, and
A.N. Venetsanopoulos,  “Vector  filtering  for  color  imaging,”
IEEE Signal Processing Magazine, vol. 22, no. 1, pp. 74–86,
2005.
[3] I. Pitas and A. Venetsanopoulos, Nonlinear Digital Filters:
Principles and Applications. Boston: Kluwer Academic
Publishers, 1990.
[4] C. G. Boncelet, “Image noise models,” in Handbook of image
and video processing, A. Bovik, Ed. Academic Press, 2000,
pp. 325–335.

© 2013 ACEEE
DOI: 01.IJSIP.4.3. 1255

48
Full Paper
ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013

[13]

[14]
[15]

[16]

[17]

[18] C. Kenney, Y. Deng, B. S. Manjunath, and G. Hewer, “Peer
group image enhancement,” IEEE Trans. on Image Processing,
vol. 10, no. 2, pp. 326–334, 2001.
[19] R. Lukac, “Adaptive color image filtering based on centerweighted vector directional filters,” Multidimensional Systems
and Signal Processing, vol. 15, no. 2, pp. 169–196, 2004.
[20] R. Lukac, B. Smolka, K. N. Plataniotis, and A.N.
Venetsanopulos, “Selection weighted vector directional filters,”
Computer Vision and Image Understanding, vol. 94, no. 1-3,
pp. 140–167, 2004.
[21] G. Peris-Fajarnes, B. Roig, and A. Vidal, “Rank-ordered
differences statistic based switching vector filter,” in Image
Analysis and Recognition, ser. Lecture Notes in Computer
Science, A. Campilho and M. Kamel, Eds. Springer-Verlag,
2006, vol. 4141, pp. 74–81.
[22] R. Lukac, B. Smolka, K. Plataniotis, and A. Venetsanopoulos,
“Vector sigma filters for noise detection and removal in color
images,” Journal of Visual Communication and Image
Representation, vol. 17, no. 1, pp. 1–26, 2006.

A.N. Venetsanopoulos, and K. Wojciechowski, “Self-adaptive
algorithm of impulsive noise reduction in color images,” Pattern
Recognition, vol. 35, no. 8, pp. 1771 – 1784, 2002.
B. Smolka and A. Chydzinski, “Fast detection and impulsive
noise removal in color images,” Real-Time Imaging, vol. 11,
no. 5-6, pp. 389–402, 2005.
R. Lukac and K. P. B. Smolka, “Sharpening vector median
filters,” Signal Processing, vol. 87, pp. 2085–2099, 2007.
B. Smolka, “Adaptive edge enhancing technique of impulsive
noise removal in color digital images,” in Proceedings of the
Third International Conference on Computational Color
Imaging, ser. CCIW’11. Springer, 2011, pp. 60–74.
B. Smolka, “Adaptive rank based impulsive noise reduction
in color images,” in IEEE International Conference on
Communications and Electronics (ICCE 2012), Hue Vietnam,
2012, pp. 355–359.
B. Smolka, “Adaptive truncated vector median filter,” in IEEE
International Conference on Computer Science and Automation
Engineering (CSAE 2012), Shanghai, China, 2011, pp. 261–
266.

© 2013 ACEEE
DOI: 01.IJSIP.4.3.1255

49

Contenu connexe

Tendances

Noise filtering
Noise filteringNoise filtering
Noise filtering
Alaa Ahmed
 
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
ijistjournal
 
APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...
APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...
APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...
International Journal of Technical Research & Application
 
Image denoising algorithms
Image denoising algorithmsImage denoising algorithms
Image denoising algorithms
Mohammad Sunny
 
Comparative study of Salt & Pepper filters and Gaussian filters
Comparative study of Salt & Pepper filters and Gaussian filtersComparative study of Salt & Pepper filters and Gaussian filters
Comparative study of Salt & Pepper filters and Gaussian filters
Ankush Srivastava
 
impulse noise filter
impulse noise filter impulse noise filter
impulse noise filter
yousef_
 
Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...
Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...
Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...
CSCJournals
 
Speckle noise reduction from medical ultrasound images using wavelet thresh
Speckle noise reduction from medical ultrasound images using wavelet threshSpeckle noise reduction from medical ultrasound images using wavelet thresh
Speckle noise reduction from medical ultrasound images using wavelet thresh
IAEME Publication
 

Tendances (20)

Noise filtering
Noise filteringNoise filtering
Noise filtering
 
IJSRDV3I40293
IJSRDV3I40293IJSRDV3I40293
IJSRDV3I40293
 
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
 
Performance Assessment of Several Filters for Removing Salt and Pepper Noise,...
Performance Assessment of Several Filters for Removing Salt and Pepper Noise,...Performance Assessment of Several Filters for Removing Salt and Pepper Noise,...
Performance Assessment of Several Filters for Removing Salt and Pepper Noise,...
 
APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...
APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...
APPRAISAL AND ANALOGY OF MODIFIED DE-NOISING AND LOCAL ADAPTIVE WAVELET IMAGE...
 
Image Restoration (Order Statistics Filters)
Image Restoration (Order Statistics Filters)Image Restoration (Order Statistics Filters)
Image Restoration (Order Statistics Filters)
 
Image denoising algorithms
Image denoising algorithmsImage denoising algorithms
Image denoising algorithms
 
Non-essentiality of Correlation between Image and Depth Map in Free Viewpoin...
Non-essentiality of Correlation between Image and Depth Map in Free Viewpoin...Non-essentiality of Correlation between Image and Depth Map in Free Viewpoin...
Non-essentiality of Correlation between Image and Depth Map in Free Viewpoin...
 
Unit3 dip
Unit3 dipUnit3 dip
Unit3 dip
 
23 an investigation on image 233 241
23 an investigation on image 233 24123 an investigation on image 233 241
23 an investigation on image 233 241
 
Comparative study of Salt & Pepper filters and Gaussian filters
Comparative study of Salt & Pepper filters and Gaussian filtersComparative study of Salt & Pepper filters and Gaussian filters
Comparative study of Salt & Pepper filters and Gaussian filters
 
Gg2411291135
Gg2411291135Gg2411291135
Gg2411291135
 
08 frequency domain filtering DIP
08 frequency domain filtering DIP08 frequency domain filtering DIP
08 frequency domain filtering DIP
 
Image denoising
Image denoisingImage denoising
Image denoising
 
impulse noise filter
impulse noise filter impulse noise filter
impulse noise filter
 
Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...
Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...
Robustness of Median Filter For Suppression of Salt and Pepper Noise (SPN) an...
 
Speckle noise reduction from medical ultrasound images using wavelet thresh
Speckle noise reduction from medical ultrasound images using wavelet threshSpeckle noise reduction from medical ultrasound images using wavelet thresh
Speckle noise reduction from medical ultrasound images using wavelet thresh
 
A NOVEL ALGORITHM FOR IMAGE DENOISING USING DT-CWT
A NOVEL ALGORITHM FOR IMAGE DENOISING USING DT-CWT A NOVEL ALGORITHM FOR IMAGE DENOISING USING DT-CWT
A NOVEL ALGORITHM FOR IMAGE DENOISING USING DT-CWT
 
Image Enhancement using Frequency Domain Filters
Image Enhancement using Frequency Domain FiltersImage Enhancement using Frequency Domain Filters
Image Enhancement using Frequency Domain Filters
 
Digital image processing techniques
Digital image processing techniquesDigital image processing techniques
Digital image processing techniques
 

En vedette

Kueh teow kung fu
Kueh teow kung fuKueh teow kung fu
Kueh teow kung fu
Mas Linda
 

En vedette (7)

Application of CAD and SLA Method in Dental Prosthesis
Application of CAD and SLA Method in Dental ProsthesisApplication of CAD and SLA Method in Dental Prosthesis
Application of CAD and SLA Method in Dental Prosthesis
 
PV Hybrid System with DSTATCOM for Residential Applications
PV Hybrid System with DSTATCOM for Residential ApplicationsPV Hybrid System with DSTATCOM for Residential Applications
PV Hybrid System with DSTATCOM for Residential Applications
 
Optimal and Power Aware BIST for Delay Testing of System-On-Chip
Optimal and Power Aware BIST for Delay Testing of System-On-ChipOptimal and Power Aware BIST for Delay Testing of System-On-Chip
Optimal and Power Aware BIST for Delay Testing of System-On-Chip
 
Finite Element Simulation and Experiment of Chip Formation Process during Hig...
Finite Element Simulation and Experiment of Chip Formation Process during Hig...Finite Element Simulation and Experiment of Chip Formation Process during Hig...
Finite Element Simulation and Experiment of Chip Formation Process during Hig...
 
Numerical Simulation of Gaseous Microflows by Lattice Boltzmann Method
Numerical Simulation of Gaseous Microflows by Lattice Boltzmann MethodNumerical Simulation of Gaseous Microflows by Lattice Boltzmann Method
Numerical Simulation of Gaseous Microflows by Lattice Boltzmann Method
 
Kueh teow kung fu
Kueh teow kung fuKueh teow kung fu
Kueh teow kung fu
 
Design and Performance Analysis of Genetic based PID-PSS with SVC in a Multi-...
Design and Performance Analysis of Genetic based PID-PSS with SVC in a Multi-...Design and Performance Analysis of Genetic based PID-PSS with SVC in a Multi-...
Design and Performance Analysis of Genetic based PID-PSS with SVC in a Multi-...
 

Similaire à Reduced Ordering Based Approach to Impulsive Noise Suppression in Color Images

PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
ijistjournal
 

Similaire à Reduced Ordering Based Approach to Impulsive Noise Suppression in Color Images (20)

IMAGE DENOISING BY MEDIAN FILTER IN WAVELET DOMAIN
IMAGE DENOISING BY MEDIAN FILTER IN WAVELET DOMAINIMAGE DENOISING BY MEDIAN FILTER IN WAVELET DOMAIN
IMAGE DENOISING BY MEDIAN FILTER IN WAVELET DOMAIN
 
Comparisons of adaptive median filter based on homogeneity level information ...
Comparisons of adaptive median filter based on homogeneity level information ...Comparisons of adaptive median filter based on homogeneity level information ...
Comparisons of adaptive median filter based on homogeneity level information ...
 
Image Denoising Using Earth Mover's Distance and Local Histograms
Image Denoising Using Earth Mover's Distance and Local HistogramsImage Denoising Using Earth Mover's Distance and Local Histograms
Image Denoising Using Earth Mover's Distance and Local Histograms
 
PID3474431
PID3474431PID3474431
PID3474431
 
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
PERFORMANCE ANALYSIS OF UNSYMMETRICAL TRIMMED MEDIAN AS DETECTOR ON IMAGE NOI...
 
Accelerated Joint Image Despeckling Algorithm in the Wavelet and Spatial Domains
Accelerated Joint Image Despeckling Algorithm in the Wavelet and Spatial DomainsAccelerated Joint Image Despeckling Algorithm in the Wavelet and Spatial Domains
Accelerated Joint Image Despeckling Algorithm in the Wavelet and Spatial Domains
 
Edge Detection with Detail Preservation for RVIN Using Adaptive Threshold Fil...
Edge Detection with Detail Preservation for RVIN Using Adaptive Threshold Fil...Edge Detection with Detail Preservation for RVIN Using Adaptive Threshold Fil...
Edge Detection with Detail Preservation for RVIN Using Adaptive Threshold Fil...
 
K010615562
K010615562K010615562
K010615562
 
Image Filtering
Image FilteringImage Filtering
Image Filtering
 
Image Filtering Using all Neighbor Directional Weighted Pixels: Optimization ...
Image Filtering Using all Neighbor Directional Weighted Pixels: Optimization ...Image Filtering Using all Neighbor Directional Weighted Pixels: Optimization ...
Image Filtering Using all Neighbor Directional Weighted Pixels: Optimization ...
 
IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...
IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...
IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...
 
IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...
IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...
IMPROVEMENT OF BM3D ALGORITHM AND EMPLOYMENT TO SATELLITE AND CFA IMAGES DENO...
 
Comparative analysis of filters and wavelet based thresholding methods for im...
Comparative analysis of filters and wavelet based thresholding methods for im...Comparative analysis of filters and wavelet based thresholding methods for im...
Comparative analysis of filters and wavelet based thresholding methods for im...
 
[IJCT-V3I2P34] Authors: Palwinder Singh
[IJCT-V3I2P34] Authors: Palwinder Singh[IJCT-V3I2P34] Authors: Palwinder Singh
[IJCT-V3I2P34] Authors: Palwinder Singh
 
Ijetcas14 479
Ijetcas14 479Ijetcas14 479
Ijetcas14 479
 
International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)
 
Study and Analysis of Impulse Noise Reduction Filters
Study and Analysis of Impulse Noise Reduction FiltersStudy and Analysis of Impulse Noise Reduction Filters
Study and Analysis of Impulse Noise Reduction Filters
 
Nonlinear Transformation Based Detection And Directional Mean Filter to Remo...
Nonlinear Transformation Based Detection And Directional  Mean Filter to Remo...Nonlinear Transformation Based Detection And Directional  Mean Filter to Remo...
Nonlinear Transformation Based Detection And Directional Mean Filter to Remo...
 
IRJET- A Review on Various Restoration Techniques in Digital Image Processing
IRJET- A Review on Various Restoration Techniques in Digital Image ProcessingIRJET- A Review on Various Restoration Techniques in Digital Image Processing
IRJET- A Review on Various Restoration Techniques in Digital Image Processing
 
IMAGE DENOISING USING HYBRID FILTER
IMAGE DENOISING USING HYBRID FILTERIMAGE DENOISING USING HYBRID FILTER
IMAGE DENOISING USING HYBRID FILTER
 

Plus de IDES Editor

Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...
Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...
Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...
IDES Editor
 
Line Losses in the 14-Bus Power System Network using UPFC
Line Losses in the 14-Bus Power System Network using UPFCLine Losses in the 14-Bus Power System Network using UPFC
Line Losses in the 14-Bus Power System Network using UPFC
IDES Editor
 
Cloud Security and Data Integrity with Client Accountability Framework
Cloud Security and Data Integrity with Client Accountability FrameworkCloud Security and Data Integrity with Client Accountability Framework
Cloud Security and Data Integrity with Client Accountability Framework
IDES Editor
 
Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...
Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...
Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...
IDES Editor
 

Plus de IDES Editor (20)

Power System State Estimation - A Review
Power System State Estimation - A ReviewPower System State Estimation - A Review
Power System State Estimation - A Review
 
Artificial Intelligence Technique based Reactive Power Planning Incorporating...
Artificial Intelligence Technique based Reactive Power Planning Incorporating...Artificial Intelligence Technique based Reactive Power Planning Incorporating...
Artificial Intelligence Technique based Reactive Power Planning Incorporating...
 
Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...
Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...
Optimal Placement of DG for Loss Reduction and Voltage Sag Mitigation in Radi...
 
Line Losses in the 14-Bus Power System Network using UPFC
Line Losses in the 14-Bus Power System Network using UPFCLine Losses in the 14-Bus Power System Network using UPFC
Line Losses in the 14-Bus Power System Network using UPFC
 
Study of Structural Behaviour of Gravity Dam with Various Features of Gallery...
Study of Structural Behaviour of Gravity Dam with Various Features of Gallery...Study of Structural Behaviour of Gravity Dam with Various Features of Gallery...
Study of Structural Behaviour of Gravity Dam with Various Features of Gallery...
 
Assessing Uncertainty of Pushover Analysis to Geometric Modeling
Assessing Uncertainty of Pushover Analysis to Geometric ModelingAssessing Uncertainty of Pushover Analysis to Geometric Modeling
Assessing Uncertainty of Pushover Analysis to Geometric Modeling
 
Secure Multi-Party Negotiation: An Analysis for Electronic Payments in Mobile...
Secure Multi-Party Negotiation: An Analysis for Electronic Payments in Mobile...Secure Multi-Party Negotiation: An Analysis for Electronic Payments in Mobile...
Secure Multi-Party Negotiation: An Analysis for Electronic Payments in Mobile...
 
Selfish Node Isolation & Incentivation using Progressive Thresholds
Selfish Node Isolation & Incentivation using Progressive ThresholdsSelfish Node Isolation & Incentivation using Progressive Thresholds
Selfish Node Isolation & Incentivation using Progressive Thresholds
 
Various OSI Layer Attacks and Countermeasure to Enhance the Performance of WS...
Various OSI Layer Attacks and Countermeasure to Enhance the Performance of WS...Various OSI Layer Attacks and Countermeasure to Enhance the Performance of WS...
Various OSI Layer Attacks and Countermeasure to Enhance the Performance of WS...
 
Responsive Parameter based an AntiWorm Approach to Prevent Wormhole Attack in...
Responsive Parameter based an AntiWorm Approach to Prevent Wormhole Attack in...Responsive Parameter based an AntiWorm Approach to Prevent Wormhole Attack in...
Responsive Parameter based an AntiWorm Approach to Prevent Wormhole Attack in...
 
Cloud Security and Data Integrity with Client Accountability Framework
Cloud Security and Data Integrity with Client Accountability FrameworkCloud Security and Data Integrity with Client Accountability Framework
Cloud Security and Data Integrity with Client Accountability Framework
 
Genetic Algorithm based Layered Detection and Defense of HTTP Botnet
Genetic Algorithm based Layered Detection and Defense of HTTP BotnetGenetic Algorithm based Layered Detection and Defense of HTTP Botnet
Genetic Algorithm based Layered Detection and Defense of HTTP Botnet
 
Enhancing Data Storage Security in Cloud Computing Through Steganography
Enhancing Data Storage Security in Cloud Computing Through SteganographyEnhancing Data Storage Security in Cloud Computing Through Steganography
Enhancing Data Storage Security in Cloud Computing Through Steganography
 
Low Energy Routing for WSN’s
Low Energy Routing for WSN’sLow Energy Routing for WSN’s
Low Energy Routing for WSN’s
 
Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...
Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...
Permutation of Pixels within the Shares of Visual Cryptography using KBRP for...
 
Rotman Lens Performance Analysis
Rotman Lens Performance AnalysisRotman Lens Performance Analysis
Rotman Lens Performance Analysis
 
Band Clustering for the Lossless Compression of AVIRIS Hyperspectral Images
Band Clustering for the Lossless Compression of AVIRIS Hyperspectral ImagesBand Clustering for the Lossless Compression of AVIRIS Hyperspectral Images
Band Clustering for the Lossless Compression of AVIRIS Hyperspectral Images
 
Microelectronic Circuit Analogous to Hydrogen Bonding Network in Active Site ...
Microelectronic Circuit Analogous to Hydrogen Bonding Network in Active Site ...Microelectronic Circuit Analogous to Hydrogen Bonding Network in Active Site ...
Microelectronic Circuit Analogous to Hydrogen Bonding Network in Active Site ...
 
Texture Unit based Monocular Real-world Scene Classification using SOM and KN...
Texture Unit based Monocular Real-world Scene Classification using SOM and KN...Texture Unit based Monocular Real-world Scene Classification using SOM and KN...
Texture Unit based Monocular Real-world Scene Classification using SOM and KN...
 
Mental Stress Evaluation using an Adaptive Model
Mental Stress Evaluation using an Adaptive ModelMental Stress Evaluation using an Adaptive Model
Mental Stress Evaluation using an Adaptive Model
 

Dernier

Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in DelhiRussian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
kauryashika82
 
1029-Danh muc Sach Giao Khoa khoi 6.pdf
1029-Danh muc Sach Giao Khoa khoi  6.pdf1029-Danh muc Sach Giao Khoa khoi  6.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
QucHHunhnh
 
Activity 01 - Artificial Culture (1).pdf
Activity 01 - Artificial Culture (1).pdfActivity 01 - Artificial Culture (1).pdf
Activity 01 - Artificial Culture (1).pdf
ciinovamais
 

Dernier (20)

Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in DelhiRussian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
Russian Escort Service in Delhi 11k Hotel Foreigner Russian Call Girls in Delhi
 
Application orientated numerical on hev.ppt
Application orientated numerical on hev.pptApplication orientated numerical on hev.ppt
Application orientated numerical on hev.ppt
 
Unit-IV; Professional Sales Representative (PSR).pptx
Unit-IV; Professional Sales Representative (PSR).pptxUnit-IV; Professional Sales Representative (PSR).pptx
Unit-IV; Professional Sales Representative (PSR).pptx
 
General Principles of Intellectual Property: Concepts of Intellectual Proper...
General Principles of Intellectual Property: Concepts of Intellectual  Proper...General Principles of Intellectual Property: Concepts of Intellectual  Proper...
General Principles of Intellectual Property: Concepts of Intellectual Proper...
 
Holdier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdfHoldier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdf
 
Mixin Classes in Odoo 17 How to Extend Models Using Mixin Classes
Mixin Classes in Odoo 17  How to Extend Models Using Mixin ClassesMixin Classes in Odoo 17  How to Extend Models Using Mixin Classes
Mixin Classes in Odoo 17 How to Extend Models Using Mixin Classes
 
On National Teacher Day, meet the 2024-25 Kenan Fellows
On National Teacher Day, meet the 2024-25 Kenan FellowsOn National Teacher Day, meet the 2024-25 Kenan Fellows
On National Teacher Day, meet the 2024-25 Kenan Fellows
 
Measures of Dispersion and Variability: Range, QD, AD and SD
Measures of Dispersion and Variability: Range, QD, AD and SDMeasures of Dispersion and Variability: Range, QD, AD and SD
Measures of Dispersion and Variability: Range, QD, AD and SD
 
Web & Social Media Analytics Previous Year Question Paper.pdf
Web & Social Media Analytics Previous Year Question Paper.pdfWeb & Social Media Analytics Previous Year Question Paper.pdf
Web & Social Media Analytics Previous Year Question Paper.pdf
 
1029-Danh muc Sach Giao Khoa khoi 6.pdf
1029-Danh muc Sach Giao Khoa khoi  6.pdf1029-Danh muc Sach Giao Khoa khoi  6.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
 
Explore beautiful and ugly buildings. Mathematics helps us create beautiful d...
Explore beautiful and ugly buildings. Mathematics helps us create beautiful d...Explore beautiful and ugly buildings. Mathematics helps us create beautiful d...
Explore beautiful and ugly buildings. Mathematics helps us create beautiful d...
 
PROCESS RECORDING FORMAT.docx
PROCESS      RECORDING        FORMAT.docxPROCESS      RECORDING        FORMAT.docx
PROCESS RECORDING FORMAT.docx
 
ICT role in 21st century education and it's challenges.
ICT role in 21st century education and it's challenges.ICT role in 21st century education and it's challenges.
ICT role in 21st century education and it's challenges.
 
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptxICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptx
 
Class 11th Physics NEET formula sheet pdf
Class 11th Physics NEET formula sheet pdfClass 11th Physics NEET formula sheet pdf
Class 11th Physics NEET formula sheet pdf
 
Activity 01 - Artificial Culture (1).pdf
Activity 01 - Artificial Culture (1).pdfActivity 01 - Artificial Culture (1).pdf
Activity 01 - Artificial Culture (1).pdf
 
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptxINDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
INDIA QUIZ 2024 RLAC DELHI UNIVERSITY.pptx
 
psychiatric nursing HISTORY COLLECTION .docx
psychiatric  nursing HISTORY  COLLECTION  .docxpsychiatric  nursing HISTORY  COLLECTION  .docx
psychiatric nursing HISTORY COLLECTION .docx
 
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
 
ComPTIA Overview | Comptia Security+ Book SY0-701
ComPTIA Overview | Comptia Security+ Book SY0-701ComPTIA Overview | Comptia Security+ Book SY0-701
ComPTIA Overview | Comptia Security+ Book SY0-701
 

Reduced Ordering Based Approach to Impulsive Noise Suppression in Color Images

  • 1. Full Paper ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013 Reduced Ordering Based Approach to Impulsive Noise Suppression in Color Images Bogdan Smolka, Krystyna Malik Silesian University of Technology, Akademicka 16, 44-100 Gliwice, Poland {Bogdan.Smolka; Krystyna.Malik}@polsl.pl Abstract—In this paper a novel filtering design intended for the impulsive noise removal in color images is presented. The described scheme utilizes the rank weighted cumulated distances between the pixels belonging to the local filtering window. The impulse detection scheme is based on the difference between the aggregated weighted distances assigned to the central pixel of the window and the minimum value, which corresponds to the rank weighted vector median. If the difference exceeds an adaptively determined threshold value, then the processed pixel is replaced by the mean of the neighboring pixels, which were found to be not corrupted, otherwise it is retained. The important feature of the described filtering framework is its ability to effectively suppress impulsive noise, while preserving fine image details. The comparison with the state-of-the-art denoising schemes revealed that the proposed filter yields better restoration results in terms of objective restoration quality measures. noise removal in color images is based on the order statistics [7]. These filters perform the vector ordering of the set of pixels from the filtering window to determine the filtering output. The widely used reduced vector ordering is based on assigning a dissimilarity measure to each color pixel from the filtering window [1,8,9]. The aggregated dissimilarity measure assigned to pixel xi is defined as (1) where denotes the distance between pixels in a given color space. The values of Ri, (i = 1,...,n) are then sorted and the vectors x1, x2, ..., xn are correspondingly ordered (2) where < denotes the order relation between vectors and R(k) denotes the k-th smallest value of R. Many denoising techniques define the vector x(1) in (2) as their output, since vectors that diverge significantly from the samples of w appear in the higher indexed locations in their ordered sequence. One of the most widely used noise reduction techniques is the Vector Median Filter (VMF), whose output is the vector x(1) from W, for which the sum of distances to all other vectors is minimized (3) Index Terms—color image processing, noise reduction, image enhancement I. INTRODUCTION Quite often the quality of color images is degraded by various types of noise, whose removal is required to enable the success of further steps in the image processing pipeline. Noise, arising from a variety of sources, is inherent to electronic image sensors and therefore the noisy signal has to be processed by a suitable filtering algorithm that reduces the noise component, while preserving original image details [1-3]. In this work we focus on a special kind of distortion, which is introduced by impulsive noise caused by malfunctioning sensors, faulty memory locations in hardware, aging of the storage material or transmission errors due to natural or man-made processes [4-6]. The color image will be treated as a two-dimensional matrix consisting of pixels , where the index u = 1,...,N indicates the pixel position on the image domain and v denotes the number of channels. In the case of standard color images v=3. The vector components xuk, for k = 1,2,3 represent the RGB color channels values. Generally, filtering operators work on the assumption that the local image features can be extracted from a small image region centered at pixel xu, called a sliding filtering window and denoted as w. Thus, the output of the filtering operation will depend only on the n samples contained within the filtering window centered at xu, which will be also denoted for convenience as , so that . The majority of the nonlinear filters used for impulsive © 2013 ACEEE DOI: 01.IJSIP.4.3. 1255 where denotes Euclidean norm. The VMF is the most popular operator intended for smoothing out the impulses injected into the color image by the noise process. This filter is very efficient at reducing the impulses, preserves sharp edges and linear trends. However, the drawback of the VMF and other filters based on vector ordering lies in introducing excessive smoothing, which results in an extensive blurring of the output images. Thus, the filters based on vector ordering do not preserve fine image structures, which are treated as noise and therefore generally show a tendency to the blurring of image details and generation of color artifacts. In order to alleviate the problem of image smoothing various switching filters, that replace only the corrupted pixels have been proposed [10-13]. The efficiency of a switching filter depends both on the quality of the impulse detection scheme and on the applied restoration framework, which replaces the detected impulses with estimates derived from the samples belonging to a local processing window. In this paper a novel switching filter is proposed. The main advantage of the proposed approach is its ability to suppress the noise component, while preserving fine image 44
  • 2. Full Paper ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013 details. The structure of the filter is based on the reduced ordering statistics and is characterized by a low computational complexity, which enables the adoption of the new technique in real-time applications. proposed method. It can be observed that the map of the detected noise correlates very well with the real contamination measured as the Euclidean distance between the original and corrupted pixels in the RGB color space normalized to the range < 0,255>. Figure 3 exhibits the correlation between the real and the detected impulsiveness of the image pixels using the test images Goldhill and Parrots. The results reveal a high correlation between the real and detected noise. II. RANK WEIGHTED FILTERING DESIGN The reduced ordering schemes are based on the sum of the dissimilarity measures between a given pixel and the samples from the local filtering window w. In this way, the output of the vector median filter is the pixel whose average distance to other pixels is minimized. The distances between the pixel xiand all other pixels w belonging to can be arranged into a sequence (4) and the ranks of the ordered distances can be used for the evaluation of the aggregated distances in (1). If stands for the rank of a given distance, then will denote the corresponding distance value and instead of the aggregated distances in (1) a weighted sum of distances, utilizing the distance ranks, can be composed (5) Figure 1. Illustration of the noise detection efficiency using the Parrots image corrupted by impulsive noise with p=0.1 intensity. where f(r) is a decreasing weighting function of the distance rank r. Then, the rank weighted sums of distances Di can be sorted and a new sequence of vectors is obtained (6) where the vector is the output of the Rank Weighted Vector Median Filter (RWVMF). Applying a step-like function (7) the Sharpening Vector Median Filter (SVMF) presented in [14] is obtained. Extensive experiments revealed that satisfactory denoising results are achieved using monotonously decreasing function f(r) = 1/r [15-17]. The weighting function decreases the influence of large distances introduced by the outliers injected by the noise process, which enables to efficiently remove the impulsive noise while enhancing the image edges. Figure 2. Illustration of the noise detection efficiency using the Goldhill image corrupted by noise with p=0.1 intensity. Figure 4 shows the histograms of the corresponding values of D1, D(1) and the measure of impulsiveness D1 - D(1) for different levels of noise intensity using the Parrots test image. As can be seen, with increasing level of noise intensity, the histograms are shifted towards higher values, which is confirmed by the mean value of D1 - D(1) marked with a red line on the plots. The structure of the proposed switching filter is quite simple. If the difference D1 - D(1), which measures the pixel corruption, exceeds a given threshold value , then a pixel is declared as corrupted by a noise process, otherwise it is treated as not disturbed III. ADAPTIVE SWITCHING FILTER In order to decide whether a pixel of a color image is corrupted by impulsive noise, the difference between the cumulated weighted distance D1 assigned to the central pixel of the filtering window and the value of D(1) corresponding to the rank weighted vector median filter output can be used [15-17] and the strength of the impulsive contamination can be estimated as the difference between D1 and D(1). Figures 1 and 2 show examples of the detected noise using parts of the color test images Parrots and Goldhill corrupted by impulsive noise, whose intensity p denotes the percentage of corrupted pixels. The visual comparison of the real and detected noise confirm the good efficiency of the © 2013 ACEEE DOI: 01.IJSIP.4.3.1255 (8) 45
  • 3. Full Paper ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013 measure D1 - D(1) computed for all image pixels and denoted as , (see also Fig. 4). The dependence between the average impulsiveness measure and noise intensity level p is shown in Fig. 7. Figure 3. Illustration of the correlation between the real noise and the output of the proposed noise detector for different levels of noise intensity. The correlation coefficient p is provided for each plot. where y1 is the switching filter output, x1 is the central pixel of the filtering window and is the Arithmetic Mean Filter output computed using only the pixels found by the detector to be not corrupted by the noise process. If all neighbors of the central pixel x1 of the filtering window are declared as corrupted, then the VMF applied to all pixels from w is taken as the filter output. Of course, the efficiency of the switching scheme is dependent mainly on the value of the thresholding parameter. If the threshold T is too low, the filter will be replacing uncorrupted pixels and much of the image details will be lost. On the other hand, if the value of T is too high, many corrupted pixels will pass the filter without being processed. As could be expected, the optimal setting of T depends on the contamination intensity. As can be observed in Fig. 6, which shows the dependence between the optimal threshold . value and the noise contamination level p evaluated using a set of test images shown in Fig. 5, the threshold yielding the best PSNR value is decreasing with increasing noise contamination level and does not depend significantly on the image structure. Therefore, the thresholding parameter T needs to be adjusted to the noise intensity level. The experiments performed using the set of images shown in Fig. 5, indicate an approximately linear dependence between the contamination ratio p and the mean value of the impulsiveness © 2013 ACEEE DOI: 01.IJSIP.4.3.1255 Figure 4. Histograms of the values of D1, D(1) and D1 - D(1) created using the PARROTS image contaminated by the impulsive noise with different intensities. The red line marks the mean value of a histogram. 46
  • 4. Full Paper ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013 Figure 5. Color test images used for the construction of an adaptive filter. Figure 1. Dependence between the optimal threshold and the output of the noise detector for the images shown in Fig. 5. IV. COMPARISON WITH E XISTING TECHNIQUES For the evaluation of the effectiveness of the proposed filtering technique the test images were corrupted by a impulsive noise modeled in such a way, that the noisy pixel is defined as: (10) Figure 6. Dependence of the optimal threshold on the contamination level p for the images shown in Fig. 5 corrupted by impulsive noise. where and , . In this impulsive noise model, the affected pixels have corrupted all three channels, which take on random values from the interval <0,255>. For the measurement of the restoration quality, the commonly used Root Mean Squared Error (RMSE) expressed through the Peak Signal to Noise Ratio (PSNR) was used, as the RMSE is a good measure of the efficiency of impulsive noise suppression and correlates well with other commonly used restoration quality measures [1]. The effectiveness of the proposed adaptive technique was compared with a set of switching filters intended for the suppression of impulsive noise in color images. The following filters were chosen for the comparison: Peer Group Filter (F1), [18],  Adaptive Center Weighted VMF (F2), [19], Adaptive Center Weighted Directional-Distance Filter (F3), [20], Rank-Ordered VMF (F4), [21], Sigma Directional Distance Filter (F5), [22]. Their parameters were set according to the recommendations provided in the appropriate references. The four images depicted in Fig. 9 were contaminated by the impulsive noise with intensities ranging from 0.05 to 0.3. It is worth noticing, that these images were not included in the set of images used to establish the dependence expressed by Eq. (9). The results are summarized in Tab. 1. As can be observed the proposed filtering method yields Figure 7. Dependence of the measure of impulsiveness on the contamination level for color images depicted in Fig. 5. Combining these plots, we obtain the scatter plot depicted in Fig. 8, which reveals a roughly linear dependence between the optimal value of the adaptive threshold and the value of (9) which can be used for the adaptive setting of the appropriate threshold value enabling high efficiency of the proposed filter for low and moderate noise contamination levels. © 2013 ACEEE DOI: 01.IJSIP.4.3. 1255 47
  • 5. Full Paper ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013 results significantly superior to those obtained using the state-of-the-art denoising methods. Figure 9. Test images used for the evaluation of the restoration results. TABLE I. COMPARISON OF THE PSNR VALUES OBTAINED WHEN RESTORING THE COLOR TEST IMAGES CONTAMINATED WITH THE IMPULSIVE N OISE USING THE PROPOSED TECHNIQUE AND SOME COMPETITIVE FILTERS Im ag e G o ld hi ll L en a Pa rr ot s Pep per s ?? NE W F1 F2 F3 F4 F5 0. 1 0. 2 0. 3 0. 1 0. 2 0. 3 0. 1 0. 2 0. 3 0. 1 0. 2 0. 3 37 .1 4 3 6. 19 3 2. 66 2 9. 14 3 8. 56 3 4. 54 3 0. 29 3 7. 32 3 3. 98 2 9. 43 3 7. 56 3 3. 42 2 8. 58 3 5. 47 3 1. 07 2 6. 29 3 7. 87 3 1. 93 2 6. 62 3 6. 73 3 1. 37 2 6. 03 3 6. 71 3 0. 79 2 5. 10 3 3. 94 3 0. 53 2 6. 84 3 6. 34 3 1. 69 2 7. 03 3 6. 78 3 1. 83 2 6. 73 3 5. 09 3 0. 27 2 4. 89 3 5. 54 3 2. 26 2 8. 52 3 8. 12 3 3. 98 2 9. 65 3 6. 29 3 3. 09 2 8. 59 3 7. 07 3 2. 75 2 7. 80 3 4. 94 2 9. 57 2 4. 13 3 6. 07 2 9. 46 2 3. 75 3 6. 70 2 9. 82 2 3. 81 3 4. 36 2 7. 59 2 1. 98 33 .8 5 31 .3 3 39 .8 1 36 .3 5 33 .3 5 38 .6 0 35 .5 5 32 .2 1 41 .6 2 38 .4 5 34 .3 1 The results summarized in Tab. 1 are confirmed by the subjective analysis of the results depicted in Fig. 10, which shows the restoration quality achieved using the proposed filter as compared with other reference filters. IV. CONCLUSIONS In this paper a new switching filtering design has been proposed. The filter is based on the weighted cumulative distances between pixels, which are used for the detection of samples corrupted by impulsive noise process. The experiments performed on test images indicate a high efficiency of the proposed filtering design. Moreover, the new filter has a low computational complexity and simple structure, which makes it attractive for real-time applications. Figure 10. Comparison of the efficiency of the proposed technique with other methods using a part of the color test image GOLDHILL contaminated by impulsive noise with intensity p = 0.1 [5] J. Zheng, K. Valavanis, and J. Gauch, “Noise removal from color images,” Journal of Intelligent and Robotic Systems, vol. 7, no. 1, pp. 257–285, August 1993. [6] H. Faraji and W.J. MacLean, “CCD noise removal in digital images,” IEEE Transactions on Image Processing, vol. 15, no. 9, pp. 2676–2685, 2006. [7] J. Astola, P. Haavisto, and Y. Neuvo, “Vector median filters,” Proceedings of the IEEE, vol. 78, no. 4, pp. 678–689, 1990. [8] B. Smolka, K. Plataniotis, and A. Venetsanopoulos, Nonlinear Signal and Image Processing: Theory, Methods, and Applications. CRC Press, 2004, chapter: Nonlinear techniques for color image processing, pp. 445–505. [9] B. Smolka and A. Venetsanopoulos, Color Image Processing: Methods and Applications. CRC Press, 2006, chapter: Noise reduction and edge detection in color images, pp. 75–100. [10] R. Lukac, “Adaptive vector median filtering,” Pattern Recognition Letters, vol. 24, no. 12, pp. 1889–1899, 2003. [11] B. Smolka, A. Chydzinski, K. W. Wojciechowski, K. N. Plataniotis, and A. N. Venetsanopoulos, “On the reduction of impulsive noise in multichannel image processing,” Optical Engineering, vol. 40, no. 6, pp. 902–908, 2001. [12] B. Smolka, K.N. Plataniotis, A. Chydzinski, M. Szczepanski, ACKNOWLEDGMENT This work was supported by the Polish National Science Center (NCN) under the Grant: DEC-2012/05/B/ST6/03428. REFERENCES [1] K. Plataniotis and A. Venetsanopoulos, Color Image Processing and Applications. Springer Verlag, 2000. [2] R. Lukac, B. Smolka, K. Martin, K. Plataniotis, and A.N. Venetsanopoulos,  “Vector  filtering  for  color  imaging,” IEEE Signal Processing Magazine, vol. 22, no. 1, pp. 74–86, 2005. [3] I. Pitas and A. Venetsanopoulos, Nonlinear Digital Filters: Principles and Applications. Boston: Kluwer Academic Publishers, 1990. [4] C. G. Boncelet, “Image noise models,” in Handbook of image and video processing, A. Bovik, Ed. Academic Press, 2000, pp. 325–335. © 2013 ACEEE DOI: 01.IJSIP.4.3. 1255 48
  • 6. Full Paper ACEEE Int. J. on Signal and Image Processing , Vol. 4, No. 3, Sept 2013 [13] [14] [15] [16] [17] [18] C. Kenney, Y. Deng, B. S. Manjunath, and G. Hewer, “Peer group image enhancement,” IEEE Trans. on Image Processing, vol. 10, no. 2, pp. 326–334, 2001. [19] R. Lukac, “Adaptive color image filtering based on centerweighted vector directional filters,” Multidimensional Systems and Signal Processing, vol. 15, no. 2, pp. 169–196, 2004. [20] R. Lukac, B. Smolka, K. N. Plataniotis, and A.N. Venetsanopulos, “Selection weighted vector directional filters,” Computer Vision and Image Understanding, vol. 94, no. 1-3, pp. 140–167, 2004. [21] G. Peris-Fajarnes, B. Roig, and A. Vidal, “Rank-ordered differences statistic based switching vector filter,” in Image Analysis and Recognition, ser. Lecture Notes in Computer Science, A. Campilho and M. Kamel, Eds. Springer-Verlag, 2006, vol. 4141, pp. 74–81. [22] R. Lukac, B. Smolka, K. Plataniotis, and A. Venetsanopoulos, “Vector sigma filters for noise detection and removal in color images,” Journal of Visual Communication and Image Representation, vol. 17, no. 1, pp. 1–26, 2006. A.N. Venetsanopoulos, and K. Wojciechowski, “Self-adaptive algorithm of impulsive noise reduction in color images,” Pattern Recognition, vol. 35, no. 8, pp. 1771 – 1784, 2002. B. Smolka and A. Chydzinski, “Fast detection and impulsive noise removal in color images,” Real-Time Imaging, vol. 11, no. 5-6, pp. 389–402, 2005. R. Lukac and K. P. B. Smolka, “Sharpening vector median filters,” Signal Processing, vol. 87, pp. 2085–2099, 2007. B. Smolka, “Adaptive edge enhancing technique of impulsive noise removal in color digital images,” in Proceedings of the Third International Conference on Computational Color Imaging, ser. CCIW’11. Springer, 2011, pp. 60–74. B. Smolka, “Adaptive rank based impulsive noise reduction in color images,” in IEEE International Conference on Communications and Electronics (ICCE 2012), Hue Vietnam, 2012, pp. 355–359. B. Smolka, “Adaptive truncated vector median filter,” in IEEE International Conference on Computer Science and Automation Engineering (CSAE 2012), Shanghai, China, 2011, pp. 261– 266. © 2013 ACEEE DOI: 01.IJSIP.4.3.1255 49