International Journal of Electronics and Communication Engineering & Technology

(IJECET)
Volume 6, Issue 9, Sep 2015, pp. 48-56, Article ID: IJECET_06_09_006
Available online at
http://www.iaeme.com/IJECETissues.asp?JTypeIJECET&VType=6&IType=9
ISSN Print: 0976-6464 and ISSN Online: 0976-6472
IAEME Publication

IMPLEMENTATION OF A NEW COLOR

FILTER ARRAY WITH IMPROVED
QUALITY FOR NOISELESS AND NOISY
COLOR IMAGES
Chaitrali Vijay Pawar
Student-ME-II Signal Processing
Dr. D.Y Patil SOEA, Ambi
Pune, India
Dr. S. D. Shirbahadurkar
Principal at Dr. D. Y. Patil COEA, Ambi
Dr. D. Y Patil SOEA, Ambi
Pune, India
ABSTRACT
Digital camera should have High sensitivity properties such as to reduce
the exposure time, to increase the aperture, or to use a lower ISO setting and
a less destructive de-noising process while acquiring a picture. The Colour
Filter Array is overlaid on the digital camera. Digital camera takes the
images by using this colour filter array. The growing popularity of digital
camera demands the improvement in the quality of the images, speed of
acquiring the images. In this paper we present the CFA with new optimal
properties which provide trade-off between robustness to aliasing,
chrominance noise and luminance noise. The proposed CFA has a natural and
simple demosaicking algorithm associated to it, inspired by its characteristics
in the Fourier domain. The demosaicking algorithm used is simple and
efficient which fully exploits the spectral properties of the CFA. The proposed
CFA work superiorly for acquisition of noiseless and noisy images and has six
colours and a periodic pattern of size 2*3. A sensor with the proposed CFA
will provide images with higher perceived resolution because the anti-alias
filter can be removed from the sensor and better quality than the standard
Bayer CFA. It is less sensitive to Inter-chrominance aliasing & chrominance
axis. The proposed CFA developed will provide less amplified chrominance
noise than Hirakawa CFA. The proposed CFA is developed which will provide
robustness in aliasing and luminance & chrominance noise.
KEY WORDS: Color filter array (CFA), Demosaicking of images,
luminance/chrominance basis, Spatio-spectral sampling.

http://www.iaeme.com/IJECET/index.asp

48

editor@iaeme.com

Implementation of A New Color Filter Array with Improved Quality For Noiseless and Noisy
Color Images

Cite this Article: Chaitrali Vijay Pawar and Dr. S. D. Shirbahadurkar.

Implementation of A New Color Filter Array With Improved Quality For
Noiseless and Noisy Color Images, International Journal of Electronics and
Communication Engineering & Technology, 6(9), 2015, pp. 48-56.
http://www.iaeme.com/IJECET/issues.asp?JTypeIJECET&VType=6&IType=9

1. INTRODUCTION
The growing popularity of digital camera demands the improvement in the quality of
the images, speed of acquiring the images. Sensor is the heart of any digital or video
camera i.e.it is 2-D array of photosites which measures the amount of light absorbed
during the exposure time. The color information is obtained by means of a color filter
array (CFA) overlaid on the sensor, such that each photosite is covered by a color
filter sensitive to only a portion of the visible light spectrum. The Colour Filter Array
is overlaid on the digital camera. Digital camera takes the images by using this colour
filter array. From the mosaicked image acquired by the camera, some processing is
required to recover a full color image with three components per pixel, carrying
information in the red (R), green (G) and blue (B) spectral bands to which the human
visual system (HVS) is sensitive. This reconstruction operation is called
demosaicking [8-10].

2. SYSTEM ANALYSIS
2.1. Existing System
The Bayer CFA, which consists of filters with the primary colors is the most popular
and dominates the consumer market. There is a vast literature dealing with the best
way to reduce aliasing artifacts during the demosaicking process, but these artifacts
are inherent to the spectral characteristics of the Bayer CFA.
Kodak patented new CFAs containing transparent (panchromatic) filters, in
addition to filters. Other CFAs have been proposed in the literature[4],[5]. However,
all these CFAs have been designed empirically and are not based on a thorough theory
for CFA design [10],[11].
A breakthrough in the field was made by Hirakawa, who proposed to design CFAs
directly in the Fourier domain, without constraints on the colors of the filters in the
spatial domain [3].
The increasing progress in CFA design and demosaicking gives minimum aliasing
artifacts due to spectral overlap of modulated colour channels in mosaicked image. By
increasing the resolution of the sensors, aliasing has become a minor issue [12],[13].
Thus robustness to the noise is more important than the robustness to the aliasing.
Digital camera should have High sensitivity properties such as to reduce the exposure
time, to increase the aperture, or to use a lower ISO setting and a less destructive denoising process while acquiring a picture. This is important for photography in lower
light level environments. Hence we need new CFA with improves sensitivity to pack
maximum energy of the colour scene into the mosaicked image.

2.2. Proposed System

The proposed CFA has a natural and simple demosaicking algorithm associated to it,
inspired by its characteristics in the Fourier domain. The demosaicking algorithm

http://www.iaeme.com/IJECET/index.asp

49

editor@iaeme.com

Chaitrali Vijay Pawar and Dr. S. D. Shirbahadurkar

used is simple and efficient which fully exploits the spectral properties of the CFA.
The proposed CFA work superiorly for acquisition of noiseless and noisy images and
has six colours and a periodic pattern of size 2*3. A sensor equipped with the
proposed CFA instead of the standard Bayer CFA will provide images with higher
perceived resolution and better quality. It is less sensitive to Inter-chrominance
aliasing & chrominance axis. The proposed CFA developed will provide less
amplified chrominance noise than Hirakawa CFA. The proposed CFA is developed
which will provide robustness in aliasing and luminance & chrominance noise.

3. DEMOSAIKING STATEGIES
3.1. Demosaiking Algorithm for the Proposed CFA
1. Compute the image v1 from v by modulation with the carrier wave of the
chrominance C
v1[k]=(-1)k1+1 2/c sin(2k2/3)v[k].
2.
3.

Apply an appropriate low-pass filter: demC1=v1*h.

Compute the image v2 from v by modulation with the carrier wave of the
chrominance C2

v2 [k] = (-1)k1 2/c cos(2k2/3)v[k].

4.

Apply the same low-pass filter: demC2=v2*h.

5.

Estimate the luminance by subtraction of the demodulated Chrominance

demL[k] =(v[k]+c(-1)k12/c sin(2k2/3) demC1 [k]c(-1)k12/c cos(2k2/3) demC2 [k])/ L

6.

Compute demR, demG , demB by pixelwise change of basis from demC1, demC2 ,demL

We remark that if h is sufficiently lowpass, the step 5) is equivalent to a convolution:

DemL =v*g

Where g[k]=( dk,0-(-1))k1 cos(2k2/3)h[k]/ L

and the Kronecker symbol is defined by
={1.. if x=y=,0..else}. This can be interesting
computationally if the luminance is to be computed in parallel with the chrominance.

The proposed algorithm can be easily adapted to handle other CFAs, simply by
using the appropriate carrier waves for the chrrominance in steps 1), 3), 5).

4. PERFORMANCE ANALYSIS
In order to compatre the performance of proposed CFA, we are going to compare
luminance and chrominance gains of the Bayer, Hirakawa and Proposed CFA.

http://www.iaeme.com/IJECET/index.asp

50

editor@iaeme.com

Implementation of A New Color Filter Array with Improved Quality For Noiseless and Noisy
Color Images

Figure 1 Luminance and chrominance gains of the Bayer, Hirakawa and proposed CFAs

Let us summarize the spectral properties of these CFAs from given fig.: In our
notations, the luminance gain is L and chrominance gain is C
TABLE Iluminance and chrominance gains
Parameters

Bayer
CFA

Hirakawa CFA

Proposed CFA

Luminance Gain : L

1/3

3/2

3/2

Chrominance Gain :
(Green Mgenta)C1
Chrominance Gain :
(Red-Blue)C2

6/4

6/4

2/4

From the comparison of three CFAs, we can say that, The Bayer CFA has high
chrominance gains, since it is made of filters with maximally saturated (primary)
colors. Hirakawas and the proposed CFAs have a higher luminance gain; they are
globally more sensitive to light. They both capture the same amount of chrominance
information: (C1)2+ (C2 )2 =1/2 But this is obtained for Hirakawas CFA at the cost
of a strong asymmetry between the two color components: C1 / C2=3; this means
that noise is more amplified in the blue-red band than in the green-magenta band
during demosaicking. Therefore Hirakawas CFA is sensitive to interchrominance
aliasing between the two chrominance bands. This may be visible in demosaicked
images at sharp color transitions for horizontally aligned objects.

http://www.iaeme.com/IJECET/index.asp

51

editor@iaeme.com

Chaitrali Vijay Pawar and Dr. S. D. Shirbahadurkar

Figure 2 Demosaiking results for three CFA sigure shows

In above figure we can see that a) Demosaicking results for the synthetic image
which consists of a sine with pulsation oscillating between green and magenta.b)With
the CFA of Hirakawa , aliasing between the two chrominance bands appears c)while
with our CFA, there is only aliasing between the chrominance and the luminance

4.1. False color artifact/color fringing

Color fringing occurs at the boundaries that separate dark and bright parts of an
image. A commonly and unfortunate artifact of Color Filter Array (CFA) demosaicing
is what is known and seen as false coloring. Typically this artifact manifests itself
along edges, where abrupt or unnatural shifts in color occur as a result of
misinterpolating across, rather than along, an edge. For preventing and removing this
false coloring various methods exist. Smooth hue transition interpolation, is used
during the demosaicing to prevent false colors from manifesting themselves in the
final image. . However, there are other algorithms that can remove false colors after
demosaicing.

4.2. Zippering artifact

The zippering artifact is another side effect of CFA demosaicing, which also occurs
primarily along edges, is known as the zipper effect. Simply put, zippering is another
name for edge blurring that occurs in an on/off pattern along an edge. This effect
occurs when the demosaicing algorithm averages pixel values over an edge, especially
in the red and blue planes, resulting in its characteristic blur. However, even with a
theoretically perfect sensor that could capture and distinguish all colors at each
photosite, Moir and other artifacts could still appear. This is an unavoidable
consequence of any system that samples an otherwise continuous signal at discrete
intervals or locations. For this reason, virtually every photographic digital sensor
incorporates something called an optical low-pass filter (OLPF) or an anti-aliasing
(AA) filter. This is typically a thin layer directly in front of the sensor, and works by
effectively blurring any potentially problematic details that are finer than the
resolution of the sensor. So in proposed CFA this color fring and zipper effect is
decreased at great extent as compare to bayer and hirakawa CFA.

http://www.iaeme.com/IJECET/index.asp

52

editor@iaeme.com

Implementation of A New Color Filter Array with Improved Quality For Noiseless and Noisy
Color Images

5. RESULTS
By comparing three CFAs, we come to know that, the Proposed Color Filter Array
provides robustness in aliasing and luminance & chrominance noise. The Proposed
Color Filter Array provides less amplified chrominance noise than Hirakawa CFA.
The Proposed Color Filter Array is less sensitive to Interchrominance aliasing &
chrominance axis. The Zipper effect and color fringes are not present in image of the
proposed CFA.
TABLE III comparison of three CFAs in terms of various parameters
Parameters

Bayer CFA

Hirakawa CFA

Proposed CFA

Typical color Fringes

Present

Not present

Not present

Zipper Effect

Present

Not present

Not present

Interchrominance Aliasing

Very sensitive

Very sensitive

Less Sensitive

Image Quality

Low

High

Very High

Luminance and Chrominance

Noise

Highly amplified

Highly amplified

Less amplified

Figure 4 Results of three CFAs

From the above discussion and from the fig 4 and fig 5, we can say that, The
Proposed Color Filter Array provide robustness inalising and luminance &
chrominance noise.

http://www.iaeme.com/IJECET/index.asp

53

editor@iaeme.com

Chaitrali Vijay Pawar and Dr. S. D. Shirbahadurkar

Figure 5.Results of three CFAs

Figure 6.Results of three CFAs

Figure 7 Results of three CFAs

http://www.iaeme.com/IJECET/index.asp

54

editor@iaeme.com

Implementation of A New Color Filter Array with Improved Quality For Noiseless and Noisy
Color Images

6. CONCLUSION
The proposed CFA has a natural and simple demosaicking algorithm associated to it,
inspired by its characteristics in the Fourier domain. The demosaicking algorithm
used is simple and efficient which fully exploits the spectral properties of the CFA.
The proposed CFA work superiorly for acquisition of noiseless and noisy images and
has six colours and a periodic pattern of size 2*3. A sensor with the proposed CFA
will provide images with higher perceived resolution because the anti-alias filter can
be removed from the sensor and better quality than that of the standard Bayer CFA.
It is less sensitive to Inter-chrominance aliasing & chrominance axis. The
proposed CFA developed will provide proposed CFA is developed which will provide
robustness in aliasing and luminance & chrominance noise and it is less sensitive to
interchrominance aliasing. In future work, we will investigate the choice of the
precise spectral sensitivity functions of the six colors defining the new CFA. The
Zipper effect and color fringes are not present in image of the proposed CFA. All the
visual results of this paper can be reproduced using the Matlab code. In future work,
we will concentrate on the choice of the precise spectral sensitivity functions of the
six colors for defining the new CFA [17].

REFERENCES
[1]

[2]
[3]
[4]

[5]

[6]
[7]
[8]

[9]
[10]
[11]
[12]

A New Color Filter Array With Optimal Properties For Noiseless And Noisy
Color Image Acquisition, Laurent Condat, Member, Ieee, Ieee Transactions On
Image Processing, 20(8), August 2011
L. Condat. A new color filter array with optimal sensing properties, presented at
the IEEE Int. Conf. Image Processing, Cairo, Egypt, Nov. 2009.
K. Hirakawa and P. J. Wolfe. Spatio-spectral color filter array design for optimal
image recovery, IEEE Trans. Image Process, 17(10), pp. 18761890, Oct. 2008.
M. Kumar, E. Morales, J. Adams, and W. Hao. New digital camerasensor
architecture for low light imaging, presented at the IEEE Int. Conf. Image
Processing, Cairo, Egypt, Nov. 2009.
Y. M. Lu and M. Vetterli. Optimal color filter array design: Quantitative
conditions and an efficient search procedure, presented at the SPIE Conf. Digital
Photography V, San Jose, CA, Jan. 2009.
L. Condat. A new random color filter array with good spectral properties,
presented at the IEEE Int. Conf. Image Processing, Cairo, Egypt, Nov. 2009.
P. Hao, Y. Li, Z. Lin, and E. Dubois, A geometric method for optimal design of
color filter arrays, IEEE Trans. Image Process, to be published.
B. K. Gunturk, J. Glotzbach, Y. Altunbasak, R. W. Schaffer, and R.M.
Mersereau, Demosaicking: Color filter array interpolation, IEEE Signal Process.
Mag 22(1), pp. 4454, Jan. 2005.
X. Li, B. K. Gunturk, and L. Zhang. Image demosaicing: A systematic survey,
presented at the IS&T/SPIE VCIP, San Jose, CA, Jan. 2008, vol. 6822.
D. Menon. Color Image Reconstruction for Digital Cameras, Ph.D.dissertation,
Univ. Padova, Padova, Italy, 2008.
K. Hirakawa and P. Wolfe, Second-generation color filter array and
demosaicking designs, presented at the SPIE VCIP, 2008, vol. 6822.
E. Dubois, Filter design for adaptive frequency-domain Bayer demosaicking, in
Proc. IEEE Int. Conf. Image Processing, Atlanta, GA, Oct. 2006, pp. 27052708.

http://www.iaeme.com/IJECET/index.asp

55

editor@iaeme.com

Chaitrali Vijay Pawar and Dr. S. D. Shirbahadurkar

[13]

[14]
[15]
[16]

[17]

[18]

[19]

B. C. de Lavarne, D. Alleysson, B. Durette, and J. Hrault, Efficient

demosaicing using recursive filtering, presented at the IEEE Int. Conf. Image
Processing, San Antonio, TX, Sep. 2007.
D. Alleysson, L. Meylan, and S. Ssstrunk, HDR CFA image rendering.
presented at the Eusipco, Sep. 2006.
A. Foi, Clipped noisy images: Heteroskedastic modeling and practical
denoising, Signal Process, 89(12), pp. 26092629, Dec. 2009.
L. Condat. A simple, fast and efficient approach to denoisaicking: Joint
demosaicking and denoising, presented at the IEEE Int. Conf. Image Processing
Hong Kong, China, 2010.
M. Parmar and S. J. Reeves. Selection of optimal spectral sensitivity functions for
color filter arrays, in Proc. IEEE Int. Conf. Image Processing, Oct. 2006, pp.
10051008.
A. Hasnat, S. Halder, A. Hoque, D. Bhattacharjee and M. Nasipuri. A Fast FPGA
Based Architecture For Measuring The Distance Between Two Color Images
Using Manhattan Distance Metric, International Journal of Electronics and
Communication Engineering & Technology, 4(3), 2015, pp. 1 - 10.
Saurabh Singh and Dr. Ashutosh Datar. An Efficient Hash Based Steganography
Technique For Text Message Using Color Images, International Journal of
Computer Engineering & Technology, 4(4), 2013, pp. 150 - 162.

http://www.iaeme.com/IJECET/index.asp

56

editor@iaeme.com