improving differently illuminant images with fuzzy membership based saturation weighting

International Journal of Advanced Engineering, Management and Science (IJAEMS) [Vol-2, Issue-6, June- 2016]
Infogain Publication (Infogainpublication.com) ISSN : 2454-1311
www.ijaems.com Page | 735
Improving Differently-Illuminant Images with
Fuzzy Membership Based Saturation Weighting
Gurpreet Kaur1
, Pooja2
, Varsha Sahni3
1.2
Department of CSE, CTIEMT, Jalandhar, Punjab, India
3
Department of Information Technology, CTIEMT, Jalandhar, Punjab, India
Abstract— Illumination estimation is basic to white
balancing digital color images and to color constancy. The
key to automatic white balancing of digital images is to
estimate precisely the color of the overall scene
illumination. Many methods for estimating the
illumination’s color has proposed. Though not the most
exact, one of the simplest and quite extensively used
methods are the gray world algorithm, white patch, max-
RGB, Gray edge using first order derivative and gray edge
using second order derivative, saturation weighting. The
first-three methods have neglected the multiple light sources
illuminate. In this work, we investigate how illuminate
estimation techniques can be improved using fuzzy
membership. The main aim of this paper is to evaluate
performance of Fuzzy Enhancement based saturation
weighting technique for different light sources (single,
multiple, indoor scene and outdoor scene) under different
conditions. The experiment has clearly shown the
effectiveness of the proposed technique over the available
methods.
Keywords— Fuzzy Enhancement, White Balancing
Illuminant, Saturation Weighting and Color Constancy.
I. INTRODUCTION
Many images, however, reveal a mixture of illuminates with
dissimilar chromaticity. Consider, such as, indoor scenes
which are affected by both indoor light sources and outdoor
light coming through the windows. So the indoor scene
shows great affectability than that of the outdoor scene and
hence the illuminate of the light sources gets buried.
Extending obtainable color constancy techniques to
effectively compute multi-illuminant estimates are a
challenging problem.
In this paper, a new method is presented that enables color
constancy under different illuminated light sources, i.e. for
single as well as for multi illuminate. As color constancy is
broadly categorized into two categories: low-level statistic
based group and learning based group. The traditional
methods like Gray World, White Patch and max-RGB all
are falling into this category, but learning based group
requires some prior knowledge.
So the proposed methodology is considered according to the
following criterion: 1) it should be able to compact with
scenes containing single and multiple light sources; 2) it
should work on a single image; 3) no human interruption is
required; 4) no prior knowledge of the light sources is
required.
In this paper, we originate and experimentally compare two
different strategies for illuminate estimation. Given an
undetected image, it is first classified as indoor or outdoor,
and then processed with the algorithm suited for that class.
• Class-Independent (CI): the same algorithm is applied
without taking into consideration the image class. The best
one has chosen by the vigorous statistical analysis.
• Class-Dependent Algorithms (CDA): for each class a
different algorithm is useful. The parameters of each
algorithm are optimized for the equivalent class. The best
algorithm for indoor and the best algorithm for outdoor are
selected by the statistical test. Given an unobserved image,
it is initially classified, and then processed with the
algorithm chosen for the predicted class.
II. COLOR CONSTANCY ALGORITHMS
Various color constancy techniques are as follows:
1. Gray world
2. White patch
3. Gray edge 1st
order derivative
4. Gray edge 2nd
order derivative
5. Modified Gray World
2.1 Gray World
Gray-World (GW) is well-known color constancy method
by the assumption which assumes that the regular
reflectance of surfaces in the world is achromatic. This
assumption is held magnificently: in an authentic world
image, in most cases it is true but presently exist lots of
different color variations. The variations colors are random

and independent; the regular would converge to the mean
value, gray, by given an enough amount of samples. Color
balancing algorithms uses this assumption by forcing its
images to have a common average gray value due to its R,
G and B components. In case, a graphic is taken by an
electronic digital camera using a particular lighting
environment, the effect from the special lighting cast is easy
to remove by enforcing the gray world assumption within
the image. As a result of approximation, along from the
image is really a lot far better an original scene [18].
2.2 White Patch
White Patch (WP) method attempts to find the objects
which have been truly white, from the scene; by assuming
the white pixels are also the brightest (I=R+G+B). White
Patch approach is conventional the Color Constancy
adaptation, trying to find the lightest patch for a white
reference comparable to how the human visual system does.
In White Patch, highest value within the image is white.
White Patch algorithm is best suited for forest category
[18].
2.3 Gray Edge 1st
order derivative
In gray Edge 1st
order derivative 4-neighbouring pixels are
considered. The primary derivative-based edge detection
operator detects image edges by computing the graphic
gradient values like Robert operator, Sobel operator, Prewitt
operator [18].
2.4 Gray Edge 2nd
order derivative
The 8-neighbouring pixels are considered, unlike 4-
connected pixels. In 8-connected, more information for
image correction is available. Gray Edge using 1st
order
derivative doesn't proof to become efficient because each
pixel considers its 4-neighbouring pixels. So, in this process
not all the information is accessible for color correction
[18].
2.5 Saturation Weighting
Saturation Weighting is based on the strong tendency of the
performance changes according to the saturation values.
Differently weighted pixels based on their saturation values
will improve the performance of the color constancy [10].
Fig.1: a) Input Image b) Result of Saturation Weighting
2.6 Modified Gray World
In this method, first, saturation weighting function is
incorporated into the gray world method, which is called as
gray world with saturation weighting (GWSW) given by:
= (1)
Where s denotes the saturation strength factor and (∙) is the
saturation weighting function,
w (f(x))= (1-S(f(x))) (2)
The smoothing operation reduces the control of noise in an
image and it was proven to be helpful for improving the
color constancy. By incorporating minowski norm (p) and
smoothing operation ( ) into GWSW, the better
performance has achieved than the general gray world
method.
III. FUZZY MEMBERSHIP BASED IMAGE
ENHANCEMENT
Previously, no work has done over fuzzy membership. This
paper work has done work on fuzzy membership based
color image enhancement using edge preserving filtering.
Fuzzy membership has used to prevent over enhancement
problem. Although, it provides enhancement to only those
objects which demands it. Also, it enhanced the object
according to membership value. Basically, it decides
whether to enhance the object or not [19].
As fuzzy membership exploits two classes because we
utilize dual membership functions in this paper work.
Improved results have extracted for color constancy from
optimal method based on fuzzy membership with edge
preservation filtering. An estimation of the proposed
technique is also drawn with existing techniques, the
comparisons have evidently shown that the fuzzy based
color constancy outperforms over the obtainable techniques
[19].
IV. PROPOSED METHODOLOGY TO IMPROVE
SATURATION WEIGHTING
Saturation Weighting is based on the strong tendency of the
performance changes according to the saturation values.
Differently weighted pixels based on their saturation values
improved the performance of the color constancy [1]. The
present paper work proposed to complete various stages
which have to be preceded following and first methodology
will discuss datasets used for dissertation work, as described
below:

concept of how far the intensity value x is from the extreme
value E.
2) The fuzzy rule for class C2 is represented as follows:
If the difference between x and E is LARGE then the
intensity of stretching should be LARGE. The above rule
indicates that the pixel values closer to E will be extensive
lesser whereas values beyond E will be extended higher
[49].
The following mathematical demonstration is used: where x
Є C2.
µD2 (X) = E − X /E – M (6)
Once the membership value for x is obtained, the contrast
enhanced xe for class C2 can be computed as follows: µD2
(X) decides what quantity of stretching parameter K and the
intensity value x has to be utilized to get the improved value
xe.
The substitution of the old x values of the V component
with the enhanced xe values will cause the V component to
be stretched resultant contrast and brightness enhanced
component Ve. This improved achromatic information Ve
can be shared with the pre-served chromatic information
(Hue and Saturation components) to obtain enhanced image
HSVe which is finally converted to enhanced RGBe image.
Fuzzy membership function
Membership function used is triangular membership
function. A triangular membership function is based on
three limits {a, b, c} as follows:
Triangle (x; a, b, c) = {0, x≤ a
x−a/ b−a, a≤ x≤ b
c−x/ c−b, b≤ x≤ c
0, c≤ x
Table.1: List of Various Parameters Used in Fuzzy
Enhancement Implementation
Nature Description
K Control parameter
M Control parameter
C1 Class one having range[0-M-1]
C2 Class two having range[M-255]
µD1 First membership value
µD2 second membership value
X intensity value
Xe Enhanced intensity value
V value
Ve Enhanced value
Phase 2: Then fuzzy membership have been added to the
saturation weighting technique in order to improve the
results.
Phase 3: But if in some cases, the edges of the image have
been loose and then in that case we have applied the edge
preserving filter as a post processing. Median filtering has
been used as an edge preserving filter. It preserves edges
while removing noise. We have made use of the 2-D
median filter where 2-D represents an array. It is useful in
preserving edges in an image. It includes further steps:
Step 1: The initial step in the proposed method is to change
the given RGB image of size P × Q into HSV along with
computing the histogram h(x) where x Є V. h(x) specifies
the number of pixels in the figure by means of intensity
value x. Proposed method uses two strengthening
parameters M and K, which handles the amount at which
the intensity value x has to be increased.
Step 2: Now, extract the V component from HSV.
Step 3: The value of K can be computed empirically
according to what level the stretching is required. From the
experimental analysis, we fixed the value 128 for K, which
gives better results for the low contrast and low bright color
images. Once the membership value of x is calculated, the
contrast enhanced value for class C2 can be computed.
Step 4: The factor M separates the histogram h(x) into two
categories or classes. The first class C1 contains pixel
values in the range [0, M − 1] and the second class C2 in
the range [M, 255].
Step 5: The stretching of V component is approved out
supported on two fuzzy membership values µD1and µD2,
calculated for C1and C2 class of pixels correspondingly.
Parameter M has the main role in the working out of fuzzy
membership values; µD1 and µD2. Enhancement parameter
K makes a decision the stretching intensity to calculate the
enhanced intensity values for the two classes C1 and C2.
Parameter K comes to a decision the stretching point to
which the intensity values x should be stretched based on
the membership values µ and µ .
Step 6: The replacement of the previous x values of the V
factor with the enhanced Values will cause the V
component to be extended resulting in contrast and
brightness enhanced component .
Step 7: This enhanced achromatic information Can be
joint with the preserved chromatic in sequence (Hue and
Saturation components) to obtain enhanced image HS
which is finally converted to enhanced RG Image.
Algorithm for fuzzy image enhancement

Edge Preserving Filter
Median filtering has used as an edge preserving filter. It
preserves edges while removing noise. We have made use
of the two-dimensional median filter where two-
dimensional represents an array. It is useful in preserving
edges in an image.
Fig.3: a) Input Image b) Modified Gray World with Edge
preserving filtering
Modified Gray World (Saturation Weighting incorporated
with gray world) which reduces the impact of the light but it
also reduces the sharpness of the image and may result in
some noise so to remove this problem; we uses an
integrated effort of the modified Gray World color
constancy with an edge preserving filtering. It not only
removes noise but also preserves edges. An edge preserving
filter is utilized to reduce the noises. The edge-preserving
filter provides extensively noise reduction.
VI. RESULT AND DISCUSSION
Performance Evaluation
Quantitative performance measures are very important in
comparing different image enhancement algorithms. Some
well-known image performance parameters for digital
images have selected to prove that the performance of the
proposed algorithm is quite better than the available
methods.
Performance is evaluated on the basis of various parameters
like MSE and RMSE. The values of the parameters are
taken in the tabular form consisting of four columns,
including image number, gray edge using first order
derivative, gray edge using second order derivative and
optimal method results.
Mean Square Error
Table 2 shows the quantized study of the mean square error.
As the mean square error needs to be reduced therefore the
algorithm is showing the better results than the available
techniques as mean square error is less in every case. The
method is experienced on various images and in each case
shows the better results than the existing method.
Table.2: MSE Evaluation
Images Gray
Edge 1
Gray
Edge 2
Proposed
Results
1. 0.0122 0.0110 0.0064
2. 0.0368 0.0307 0.0177
3. 0.0299 0.0287 0.0208
4. 0.0329 0.0283 0.0106
5. 0.0521 0.0469 0.0245
6 0.0550 0.0509 0.0283
7 0.0412 0.0374 0.0291
8 0.0080 0.0076 0.0050
9 0.0204 0.0194 0.0131
10. 0.0247 0.0224 0.0141
Graph 1: MSE Evaluation
Root Mean Square Error
Table 3 is showing the relative analysis of the root mean
square error. Table3 has evidently shown that is root mean
square error is less in our case therefore the algorithm has
shown better results over the available algorithm.
for counter = 1:length(low)
for index =
low(counter):high(counter)
transformationMap(index) =
round((low(counter)-1) +
(range(counter)*(sum(fHistogram(low
(counter):index)))/(sum(fHistogram(
low(counter):high(counter))))));
end
end
fuzzy_factor_image =
stretchlim(ip_image, fuzzy_factor);
final_image = imadjust(ip_image,
fuzzy_factor_image, []);

Table.3: RMSE Evaluation
Images Gray
Edge 1
Gray
Edge 2
Proposed
Results
1. 0.1105 0.1050 0.0803
2. 0.1919 0.1753 0.1330
3. 0.1729 0.1695 0.1441
4. 0.1814 0.1683 0.1028
5. 0.2281 0.2164 0.1564
6. 0.2346 0.2255 0.1683
7. 0.2030 0.1935 0.1707
8. 0.0892 0.0874 0.0749
9. 0.1429 0.1391 0.1146
10. 0.1573 0.1497 0.1186
Graph 2 shows the quantized analysis of the Root mean
squared Error of different images. It is very clear from the
graph that there is decrease in RMSE value of images with
the use of proposed method over existing method.
Graph 2: RMSE Evaluation
Peak Signal to Noise Ratio
Table 4 shows the comparative analysis of the Peak Signal
to Noise Ratio (PSNR). As PSNR need to be maximized; so
the main goal is to increase the PSNR. Table 4 has clearly
shown that the PSNR is maximum in all the values;
therefore algorithm is providing better results than the
available techniques. The method is tested on various
images and in each case shows better results than the
existing method.
Table.4: PSNR Evaluation
Images Gray
Edge 1
Gray
Edge 2
Proposed
Results
1. 64.8019 67.7085 70.0371
2. 62.4702 63.2560 65.6552
3. 63.3729 63.5477 64.4637
4. 62.9601 63.6105 67.8928
5. 60.9665 61.4237 64.2464
6. 60.7252 61.0669 63.6075
7. 61.9805 62.3981 63.4861
8. 69.1239 69.3033 70.6386
9. 65.0285 65.2618 66.9445
10. 64.1981 64.6286 66.6473
Graph 3 shows the quantized study of the peak signal to
noise ratio of different images. It is clear from the plot that
there is raise in the PSNR value of images with the use of
method over existing methods. This increase represents an
enhancement in the objective quality of the image.
Graph 3: PSNR
VII. CONCLUSION
To conclude, the methodology in this paper has shown to
extend the existing methods under that scenario where the
uniform light-source assumption is too restrictive. In this
paper, a new methodology that can be used to apply color
constancy for the images that are recorded in the presence
of different light sources has been proposed. Several
traditional techniques such as Grey-world technique, Max
RGB and learning-based technique were utilized to check
the color constancy of digital images suffering from a light
source. Every one of these techniques has an evident
drawback that the source of light throughout the scene is
spectrally same.
This assumption is usually violated as there can be multiple
source of light illuminating the scene. More focuses on
changing the saturation weighting based color constancy
using fuzzy membership based color image enhancement
and edge preserving filtering.

The issue appears to be justifiable and has great effect
vision application since as fuzzy membership based
saturation weighting that has decrease the impact of the
light but it additionally decreases the sharpness of the image
and also has result in certain noise so to eliminate this issue,
an integral effort of the edge based color constancy
combined with histogram stretching and edge preserving
filtering.So the proposed methodology is considered
according to the following criterion: 1) it should be able to
compact with scenes containing single and multiple light
sources; 2) it should work on a single image; 3) no human
interruption is required; 4) no prior knowledge of the light
sources is required. This scenario proves helpful for indoor
as well as outdoor illuminants.
This paper work has not considered any soft computing
technique for color constancy to check the optimistic value
of a light source for differently-illuminant images. Also the
effect of the noise is also ignored, therefore in near future,
some well-known image filters will use which depicts the
light source for differently-illuminant images under
different scenario.
REFERENCES
[1] Ahn, Hyunchan, Soobin Lee, and Hwang Soo Lee
Improving color constancy by saturation weighting, In
proceedings of IEEE International Conference on
Speech and Signal Processing (ICASSP), pp.1909-
1913, 2013.
[2] Arjan Gijsenij and Theo Gevers Color Constancy
Using Natural Image Statistics and Scene Semantics,
In proceedings of IEEE Conference on Pattern
Analysis and Machine Intelligence, Vol-33(4), 2011.
[3] Arjan Gijsenij, Gevers, Weijer, “Computational Color
Constancy: Survey and Experiments” in proceedings
of IEEE Conference on image processing, Vol-10(10),
2010.
[4] Abdeldjalil Madi, Djemel Ziou, and Frederic
Dhalleine "Exploiting color constancy for
compensating projected images on non-white light
projection screen", in proceedings of IEEE
International Conference on Computer and Robot
Vision (CRV), pp. 210-217, 2013.
[5] Bleier, Michael Color constancy and non-uniform
illumination: Can existing algorithms work, In
Computer Vision Workshops (ICCV Workshops), pp.
774-781, 2011.
[6] Bianco, Simone, and Raimondo Schettini Color
constancy using faces, In proceedings of IEEE
International Conference on Computer Vision and
Pattern Recognition (CVPR), pp. 65-72, 2012.
[7] Brown, Lisa, Ankur Datta, and Sharathchandra
Pankanti Exploiting Color Strength to Improve Color
Correction In IEEE International Symposium on
Multimedia (ISM), 2012.
[8] Bianco, S., and R. Schettini Computational color
constancy IEEE 3rd European Workshop on Visual
Information Processing (EUVIP), pp. 1-7, 2011.
[9] Cepeda-Negrete, Jonathan, and Raul E. Sanchez-
Yanez Combining color constancy and gamma
correction for image enhancement, In proceedings of
IEEE conference on Electronics, Robotics and
Automotive Mechanics Conference (CERMA), pp. 25-
30, 2012.
[10]Chang, Feng-Ju, and Soo-Chang Pei Color constancy
via chromaticity neutralization: From single to
multiple illuminants IEEE International Symposium
on Circuits and Systems (ISCAS), pp. 2808-2811,
2013.
[11]Catarina Barata, M. Emre Improving Dermoscopy
Image Classification Using Color Constancy IEEE
Journal of Biomedical and Health Informatics,Vol-
19(3), pp. 1146-1152, 2014.
[12]Chakrabarti, Ayan, Keigo Hirakawa, and Todd Zickler
Color constancy with spatio-spectral statistics IEEE
Transactions on Pattern Analysis and Machine
Intelligence, Vol-34(8), pp.1509-1519, 2012.
[13]Ebner, Marc, German Tischler, and Jürgen Albert
Integrating color constancy into JPEG2000 IEEE
Transactions on Image Processing, Vol-16(11), pp.
2697-2706, 2007.
[14]Finlayson, Graham, and Steven Hordley Improving
gamut mapping color constancy IEEE Transactions
on Image Processing, Vol-9(10), pp. 1774-1783, 2000.
[15]Gijsenij, Arjan, Rui Lu, and Theo Gevers Color
constancy for multiple light sources IEEE
697-707, 2011.
[16]Gijsenij, Arjan, Theo Gevers, and Joost Van De
Weijer Improving color constancy by photometric
edge weighting IEEE Transactions on Pattern
Analysis and Machine Intelligence, Vol-34(5) , pp.
918-929, 2012.
[17]Graham D. Finlayson, Hordley and Paul Color By
Correlation: A Simple, Unifying Framework for Color
Constancy, In proceedings of IEEE Conference on

pattern Analysis and Machine Intelligence, Vol-
23(11), pp. 1209-1221, 2001.
[18]Gurpreet Kaur and Pooja Color Constancy
Algorithms: Significant Study, In proceedings of
International Conference on Communication,
Information and Computing Technology (ICCICT-15),
pp. 690-694, 2015.
[19]Gurpreet Kaur and Pooja Improving Saturation
Weighting Color Constancy with Fuzzy Membership
and Edge Preservation An International Journal Of
Engineering Sciences , Vol-16(1), 2015.
[20]Gurpreet Kaur and Pooja Optimal CC Method:
Improved Results International Journal of Signal
Processing, Image Processing and Pattern
Recognition (IJISIP), 2016 (in press).
[21]Joze, Hamid Reza Vaezi, and Mark S. Drew
Exemplar-Based Color Constancy and Multiple
Illumination IEEE Transactions on Pattern Analysis
and Machine Intelligence, Vol-36(5), pp. 860-873,
2013.
[22]Joze, Vaezi, Hamid Reza, and Mark S. Drew White
patch gamut mapping color constancy, In proceedings
of IEEE International Conference on Image
Processing (ICIP), pp. 801-804, 2012.
[23]Joost van de Weijer and Elli Angelopoulou Multi-
Illuminant Estimation with Conditional Random
Fields IEEE Transactions on Image Processing, Vol-
23(1), pp. 83-96, 2013.
[24]J. Zhu, K. Samuel, S. Masood, and M. Tappen
Learning to Recognize Shadows in Monochromatic
Natural Images, In proceedings of IEEE Conference
on Computer Vision and Pattern Recognition, pp. 223-
230, 2010.
[25]Kobus Barnard, Adam Coath and Brian Funt A
Comparison of Computational Color Constancy
Algorithms—Part II: Experiments With Image Data,
In proceedings of IEEE Conference on image
processing, Vol-11(9), pp. 985-996, 2002.
[26]K. Saenko, B. Kulis, M. Fritz, and T. Darrell Adapting
Visual Category Models to New Domains, In
proceedings of European Conference on Computer
Vision, pp. 213-226, 2010.
[27]Kai-Fu Yang, Shao-Bing Gao and Yong-Jie Efficient
Illuminant Estimation for Color Constancy Using Grey
Pixels, In proceedings of IEEE conference on
computer vision and pattern recognition (CVPR), pp.
2254-2263, 2015.
[28]Lee, Woo-Ram, Dong-Guk Hwang, and Byoung-Min
Jun Comparison of color constancy methods for skin
color under colored illuminants, In proceedings of
IEEE International Conference on Digital Content,
Multimedia Technology and its Applications (IDCTA),
pp. 80-83, 2011.
[29]Li, Bing, Weihua Xiong, Weiming Hu, and Ou Wu
Evaluating combinational color constancy methods on
real-world images, In proceedings of IEEE Conference
on In Computer Vision and Pattern Recognition
(CVPR), pp. 1929-1936, 2011.
[30]Lu, Rui, Arjan Gijsenij, Theo Gevers, Koen EA van de
Sande, Jan-Mark Geusebroek, and De Xu Color
constancy using stage classification, In proceedings of
IEEE International Conference on Image Processing
(ICIP), pp. 685-688, 2009.
[31]Lu, Rui, Arjan Gijsenij, Theo Gevers, Vladimir
Nedovic, De Xu, and J-M. Geusebroek Color
constancy using 3D scene geometry, In proceedings of
IEEE International Conference on Computer Vision,
pp. 1749-1756, 2009.
[32]L. Shi, W. Xiong, and B. Funt Illumination Estimation
via Thin- Plate Spline Interpolation Journal of the
Optical Society of America, Vol-28(5), pp. 940-948,
2011.
[33]Madi, Abdeldjalil, and Djemel Ziou Color constancy
for visual compensation of projector displayed
image Displays, Vol-35(1), pp. 6-17, 2014.
[34]Monari, Eduardo Color Constancy Using Shadow-
Based Illumination Maps for Appearance-Based
Person Re-identification, In proceedings of IEEE
International Conference on Advanced Video and
Signal-Based Surveillance (AVSS), pp. 197-202, 2012 .
[35]Mathew, Alex, Ann Theja Alex, and Vijayan K. Asari
A manifold based methodology for color constancy, In
proceedings of IEEE Applied Imagery Pattern
Recognition Workshop (AIPR), pp.1-7, 2010.
[36]Moreno, Ramon, Manuel Grana, and Alicia D Anjou
An image color gradient preserving color constancy in
computational intelligence group, 2010.
[37]M. S. Drew, H. R. V. Joze, and G. D. Finlayson The
zeta-image, illuminant estimation, and specularity
manipulation Computer Vision and Image
Understanding, Vol-127, pp.1-13, 2014.
[38]N Banic and S. Loncaric Improving the white patch
method by sub-sampling, In proceedings of IEEE
International Conference on Image Processing (ICIP),
pp. 605-609, 2014.
[39]Nikola Banic and Sven Loncaric Color Cat:
Remembering Colors for Illumination Estimation

IEEE Signal Processing Letters, Vol-22(6), pp. 651-
655, 2014.
[40]Rezagholizadeh, Mehdi, and James J. Clark Edge-
based and Efficient Chromaticity Spatio-Spectral
Models for Color Constancy, In proceedings of IEEE
International Conference on Computer and Robot
Vision (CRV), pp.188-195, 2013.
[41]Shengxian, Cao, Du Bangkui, Sun Jiawei, Liu Fan,
Yang Shanrang, and Xu Zhiming A colour constancy
algorithm based on neural network and application
World Congress on Intelligent Control and
Automation (WCICA), pp.3100-3103, 2010.
[42]Shaobing Gao, Kaifu Yang, Chaoyi Li, Yongjie Li A
Color Constancy Model with Double-Opponency
Mechanisms IEEE Xplore, pp.929-936, 2013.
[43]S. Bianco, G. Ciocca, C. Cusano, and R. Schettini
Automatic color constancy algorithm selection and
combination Pattern recognition, Vol-43(3), pp. 695-
705, 2010.
[44]S. Bianco and R. Schettini Adaptive color constancy
using faces IEEE Transactions on Pattern Analysis
and Machine Intelligence, Vol-36(8), pp. 1505-1518,
2014.
[45]S. Gao, W. Han, K. Yang, C. Li, and Y. Li Efficient
color constancy with local surface reflectance
statistics, In proceedings of European Conference on
Computer Vision, pp. 158-173, 2014.
[46]T. Gevers, A. Gijsenij and J. van de Weijer
Generalized Gamut Mapping Using Image Derivative
Structures for Color Constancy International Journal
of Computer Vision, Vol-86(2), pp.127-139, 2010.
[47]Teng, SJ Jerome Robust Algorithm for Computational
Color Constancy IEEE International Conference on
Technologies and Applications of Artificial
Intelligence (TAAI), pp. 1-8, 2010.
[48]Tara Akhavan A new combining learning method for
color constancy, In proceedings of IEEE International
Conference on Image Processing Theory Tools and
Applications (IPTA), pp.421-425, 2010.
[49]G. Raju and Madhu S. Nair A fast and efficient color
image enhancement method based on fuzzy-logic and
histogram ELSEVIER, Vol-68(3), pp. 237-243, 2014.
[50]Van De Weijer, Joost, Theo Gevers, and Arjan
Gijsenij Edge-based color constancy IEEE
2207-2214, 2007.
[51]Vazquez-Corral, Javier Color constancy by category
correlation IEEE Transactions on Image Processing,
Vol-21(4), pp. 1997-2007, 2012.
[52]Wu, Meng, Jun Zhou, Jun Sun, and Gengjian Xue
Texture-based color constancy using local regression,
In proceedings of IEEE International Conference
on Image Processing (ICIP), pp. 1353-1356, 2010.
[53]https://in.mathworks.com/matlabcentral/fileexchange/
51959-color-constany accessed on Jul. 4,2015.
[54]Yu, Jing, and Qingmin Liao Color Constancy-Based
Visibility Enhancement in Low-Light Conditions, In
Digital Image Computing: Techniques and
Applications (DICTA), pp.441-446, 2010.
[55]http://www.cs.sfu.ca/~colour/research/colour-
constancy.html accessed on May 3, 2015.
[56]Y. Cao and A. Bermak An analog gamma correction
method for high dynamic range applications IEEE
International SOC Conference (SOCC), pp.318-322,
2011.
[57]Vaezi Joze, H, and M. Drew Exemplar-Based Color
Constancy and Multiple Illumination IEEE
Transactions on Pattern Analysis and Machine
Intelligence, Vol-36(5), pp. 860-873, 2013.

improving differently illuminant images with fuzzy membership based saturation weighting

More Related Content

What's hot (19)

Similar to improving differently illuminant images with fuzzy membership based saturation weighting (20)

Recently uploaded (20)

improving differently illuminant images with fuzzy membership based saturation weighting