International Conference of Advance Research and Innovation (ICARI-2020)

Analysis of CCD and CMOS Sensor Based Images from Technical and
Photographic Aspects
Kamaljeet Kainth1, Baljeet Singh2
1
Dept. of Electronics and Communication, GNA University, Phagwara, Punjab, India.
2
Dept. of Animation and Multimedia, GNA University, Phagwara. Punjab, India
Abstract: In Image capturing devices, Image sensors play a very vital role. Image sensor is responsible for capturing light from outside world
and converts it into electronics signal to develop a digital Image. There are two majorly used Image sensors CCD and CMOS. Both have their
own merits and demerits. When it comes to study the features of these sensors, there is always confusion whether one person goes for technical
aspects or image quality delivered by cameras based on these sensors. This paper will evaluate the both technical and photographic aspects.
Firstly, distinct images are captured by CCD and CMOS sensor based cameras by varying different parameters like ISO, exposure value, shutter
speed etc. Secondly, to test the technological aspects Mean Square Error (MSE), Pixel to Signal Noise Ratio (PSNR) and SSIM are evaluated for
different cases.
Keywords: ISO, MSE, Exposure Value, PSNR, Shutter Speed, SSIM, CCD and CMOS.

1. Introduction like ISO, Exposure value, shutter speed etc., this paper will can
Digital cameras have brought a revolution in the field of photography. facilitate a clear picture of the same. In addition to this, above
It became very easy and compatible to capture photographs or images questions considered for bot CCD and CMOS image
through digital cameras and without bothering about negative films sensors.Therefore, this paper is organized in following sections:
like film cameras. This is an easy approach to capture the photographs Section I: It covers the basic introduction as composed above.
and assure the output (digital image) instantly [1]. The significant Section II: This section will clearly define the methodology adopted
technology behind the digital cameras whic1h captures the image is in evaluating various image parameters w.r.t CMOS and CCD sensor
associated with Image Sensors. Image sensors catch the photograph based Images.
which is the case of a film camera is done by the film negative. Image Section III: Section three is dedicated to the experiment and analysis
sensors are the electronic components that capture the light in the of captured by CCD and CMOSbased cameras. This is followed by
form of information which is further converted into electronic signals the images captured by CMOS sensor based cameras alone. The
to make an image that results in the final digital image that was technical parameters are calculated w.r.t the variation in photographic
captured through our digital camera [2]. aspects. This is the major essence of this paper.
In the world of digital cameras there are majorly two types of image Section IV:This section discussed the results in details for the cases
sensors i.e. CCD (Charged coupled device) and CMOS mentioned above.
(complementary metal-oxide-semiconductor). Both image sensors The conclusion has been complied within the finish of this
convert light signals and electrons to the electronic data which is paper.
further converted in the form of image, but they of converting the 2. Methodology
light signals to readable data is different [2]. CCD (Charged coupled To achieve the above objectives, it is the vital thing to understand the
device) Image sensor signal isshifted across the electronic chip and is concepts related to image processing and capturing. This lead to the
acquire by the amplifier at one corner [3][4].With this process, CCD development of the various scenario which can fulfill the
image sensors are capable of transporting the signalsacross the chip objectives.Fig.1 depicts the steps involved in acquiring and analyzing
without any distortion or loss in the signals. the images. Analysis of the images taken by CCD and CMOS is
CCD image sensors are manufactured by the process which leads to measured on the basis of three image analysis parameters, are :
high-quality sensors in terms of light-sensitive and image quality.
1. Capture the image of an object using CCD sensor
CCD Sensors are used in high end cameras used for broadcasting or
capturing high-quality images. These images sensors produce high
based camera.
quality images with less noise. On the other hand, CMOS 2. Capture the image of an object using CMOS sensor
(Complementary metal-oxide semiconductor) signal is acquired at based camera.
each pixel of the image sensor and then the amplified signals 3. Analysis of CMOS and CCD based Images.
transferred using the traditional circuit [5]. With this process, the 4. If no difference is observed conclusion and final
CMOS approach is more flexible because each pixel can be read results.
individually to get the data and convert it to an image. By this
manufacturing process, CMOS image sensors can access signals at
else go to step 5
much high speed as compared to CCD [6]. CMOS sensors are more 5. Analysis of step 1 and step 2 images by varying ISO
used in the still photography cameras and where the cost of the and Exposure values.
equipment also has to keep low. In comparison to the CCD images 6. Capture images using recent CMOS based cameras.
sensor CMOS image sensor has less power consumption, faster data 7. Analysis of images captured at step 6 by varying ISO
read and write speed. Manufacturing is very easy and cheap in Levels, Exposure values and Shutter Speed.
comparison to the CCD [7-8].The evaluation of manufacturing
8. Evaluation of MSE, PSNR and SSIM results for step
process and comparison of two sensors on the basis of other
parameters is far away from the scope of this paper. This paper will 5 and step 6 images.
try to give answers to the readers who want understand the 9. Evaluation of MSE, PSNR and SSIM results for step
engineering (referred as technical in this paper) and photography 7.
aspects. The technical aspect discussed in this paper are Mean Square 10. Final conclusion.
Error, Pixel to Noise Ratio (PSNR), and Structure Similarity Index a) Mean Square Error, b) Pixel to Signal Noise Ratio and c) Structure
(SSIM). These parameters [9-10] are considered to any image after Similarity Index (SSIM)[11-12]. In addition to this, further analysis
some analysis. Secondly, if a photographer curious to understand is compiled by varying the ISO and Exposure values.
what will happen to an image if one’s various parameter Firstly, the images are captured using two distinct cameras. One is
having CMOS sensor and another is of having CMOS sensor. The
*Corresponding Author, resolution of images is 4068 (rows) *3072 (column), but to display in
E-mail address: baljit.singh@gnauniversity.edu.in this paper, images are scaled to lower resolution. Fig.2. shows the
All rights reserved: http://www.ijari.org images captured by CMOS and CCD based cameras.
Further, for the same object, images are captured by varying ISO
levels as ISO100, ISO200, ISO800, ISO1600, ISO3200 and Exposure

IJARI
14
Electronic copy available at: https://ssrn.com/abstract=3559236
 International Conference of Advance Research and Innovation (ICARI-2020)
values of -2, -1,0,1, and 2 Fig.3& Fig.4 depicts the images for distinct
ISO and exposure values. 100
III.As mentioned above, to observe the effect of variation on the
image, distinct image parameters are evaluated as:
a) Mean Square Error (MSE): Mean squared error is the
squared error between the reference image and the targeted
image. It results into a non-negative value. It is calculated by
200
following mathematical expression:
𝑴𝑺𝑬 = ∑𝑴,𝑵(𝑰 − 𝑰′ ).𝟐 /(𝑴 ∗ 𝑵) …………… (1)
Closer the value to zero represent less error in both
images. Where M, N are the number of rows and columns. “I”
represents the original/referenced image and “ I’ ” depicts the
targeted image[13][14]. 400
b) Pixel to Signal Noise Ratio (PSNR): PSNR is calculated with
the help of MSE, and it is defined as the ratio of the maximum
power of signal (pixel value in case of images) to the noise that
distort the quality of an image. It is represented in dB and
higher the vallue of PSNR, good is the quality of the image. 800
The mathematical expression for the PSNR is given as :
PSNR = 10*log10(𝟐𝟓𝟓)𝟐 / (𝑴𝑺𝑬)𝟐 ................(2)
Where MSE is the value of Mean Square Error calculated in
previous step and 255 is value of maximum pixel value in an
RGB image[15]. 1600
c) Structural Similarity Index(SSIM): It is a method method for
auspicating the quality of an image apfte processing. In this
paremter, SSIM is caluclated by using inbuilt function of Fig. 3: Images captured by CCD and CMOS sensor based cameras
MATLAB. The mathematial expression of the same is as: with different ISO levels
ssimval = ssim(A,ref) .......................(3) Exposure CCD sensor based CMOS sensor based
Where ref is reffred to as refrence image and A refed as the value Image Image
other Image[16-17].
As nowadays CMOS sensor-based parameters are utilized for -2
photography[8]. Hence, to analyze the performance of images which
are captured by CMOS sensor-based camera, images are acquired on
different ISO levels,exposure values, and shutter speed. The ISO and
Exposure value are same as of CCD sensor based cameras. The
shutter speed is varied as 1/50, 1/40, 1/30, 1/25 and 1/20. Fig.5 depicts
the images for distinct ISO and exposure values and Shutter speed of -1
CMOS sensor-based images. In addition to the above step, MSE,
PSNR, and SSIM are also calculated by using equation 1, 2, 3 for
CMOS sensor-basedimages. On the basis of results obtained from
MSE, PSNR, SSIM a comparison and conclusion to be made within 0
finish of the paper.
Results calculated for MSE, PSNR and SSIM against CCD-CMOS
and CMOS alone, sensor based cameras are described and composed.
4. Experiment and Analysis
4.1Analysis of image captured by CCD and CMOS 1
As discussed in methodology section, firstly different images are
captured by two distinct cameras having CCD and CMOS sensors.
To analyses the image performance, mean square error between two
images (CCD and CMOS) is calculated by using equation number 1.
Further, pixel to signal noise ratio is calculated by using equation 2 2
and SSIM using equation 3 Table 1 shows the corresponding values
of MSE, PSNR and SSIM between CMOS and CCD Images.
4.2 Analysis of CCD and CMOS images by varying ISO Fig. 4: Images captured by CCD and CMOS sensor based cameras
To analyses the image performance further, images from both CCD with different Exposure Values
and CMOS sensor based cameras are captured and different ISO
levels. The ISO levels are varied as ISO100, ISO200, ISO400,
ISO800 and ISO1600. On the basis of equation 1, 2 and 3 the MSE,
PSNR and SSIM are calculated and depicted in Table 2.

a) b)
Fig. 2. a) CCD sensor based Image; b) CMOS sensor based Image
ISO CCD sensor based CMOS sensor based
Value Image Image
Fig.5: Images captured by CMOS sensor based cameras with
different ISO and Exposure Values

IJARI 15

Electronic copy available at: https://ssrn.com/abstract=3559236

 International Conference of Advance Research and Innovation (ICARI-2020)
Table 4: MSE, PSNR and SSIM values by varying ISO Values of
ISO VALUE CMOS Sensor Based Image CMOS alone
100 ISO MSE PSNR SSIM
100-200 0.000054 90.79 0.5735
100-400 0.000017 85.82 0.1871
200 100-800 0.000017 85.82 0.1354
100-1600 0.000017 85.82 0.0545
Table 5: MSE, PSNR and SSIM values by varying Shutter Speed of
. CMOS alone
400 Shutter MSE PSNR SSIM
Speed
1/50-1/40 0.000003 103.58 0.5029
1/50-1/30 0.000011 97.56 0.5317
800 1/50-1/25 0.000023 94.58 0.5222
1/50-1/20 0.000048 91.35 0.4991
4.3Analysis of CCD and CMOS images by varying
1600 Exposure Value
To analyses the image performance further, images from both CCD
and CMOS sensor based cameras are captured and with different
exposure values. The exposure values are varied as -2, -1,0,1,2. On
Shutter Speed CMOS Sensor Based Image the basis of equation 1,2 and 3 the MSE, PSNR and SSIM are
calculated and depicted in Table 3.
4.4Analysis of Images Captured by CMOS sensor based Cameras
1/50 As per recent technology most of the cameras has CMOS sensors. To
analyse the image parameters of CMOS captured images, distinct
images are taken at different ISO levels and exposure values. Further,
MSE, PSNR and SSIM is analysed in MATLAB. For the comparison,
in case of ISO levels images with ISO100 is taken as reference images
1/40
w.r.t to the images of ISO200,400,800 and 1600 levels. The value for
MSE, PSNR and SSIM using equations are depicted in Table4.
In addition to this images are captured on distinct exposure values.
1/30 The exposure value are exposure values are -2, -1,0,1 and 2
respectively. Further, the images are capture at distinct shutter speed
of the images. The shutter speed varies as 1/50,1/40,1/30,1/25 and
1/20. Here, the image captured at a shutter speed of 1/50 is considered
as the reference image w.r.t other shutter speed. Therefore, MSE,
1/25 PSNR and SSIM are calculated between reference and target image
at different exposure values. Table 5 shows MSE, PSNR and SSIM
values by varying Shutter Speed.
CCD VS CMOS-MSE
CCD vs CMOS MSE
1/20
0.0002

Table 1.: MSE, PSNR and SSIM values between CCD and CMOS
0.0001
captured Images
S.NO Parameter Value
1 MSE 0.000042 0
2 PSNR 91.95 100 200 400 800 1600
3 SSIM 0.000992 a)
CCD VS CMOS-PSNR
Table 2.MSE, PSNR and SSIM values by varying ISO levels
PSNR
ISO MSE PSNR SSIM 150
100 0.000001 124.41 0.2322
200 0.000001 124.41 0.1778 100
400 0.000003 102.83 0.0999
800 0.000064 90.09 0.0969 50
1600 0.000151 86.35 0.0947
Table 3.Analysis of CCD and CMOS images by varying Exposure 0
Value 100 200 400 800 1600
Exposure MSE PSNR SSIM b)
Value
-2 0.000044 91.74 0.1518
-1 0.000029 93.53 0.1497
0 0.000002 104.41 0.2621
1 0.000002 104.41 0.2413
2 0.000000 112.37 0.1646

IJARI 16

Electronic copy available at: https://ssrn.com/abstract=3559236

 International Conference of Advance Research and Innovation (ICARI-2020)

CCD VS CMOS-SSIM decreases with decrease in the value of SSIM.Fig 6 shows the
variation in MSE, PSNR and SSIM values on different ISO values.
SSIM Fig 7 shows the variation in MSE, PSNR and SSIM for different
exposure values. On the other hand, for CMOS based photographs
0.25
the quality of images also decreases as the ISO level increases, which
0.2 means it has higher value of MSE and hence corresponding trend of
PSNR.
0.15
CMOS ISO (REF 100 ISO)-MSE
0.1
MSE ( with ref to 100 ISO)
0.05
0 0.0001
100 200 400 800 1600
0.00005
c)
Fig.6: The variation in MSE, PSNR and SSIM values on different 0
ISO values for CMOS & CCD Images A B C D
5. Results and Discussions a)
To explore the image quality captured by CMOS and CCD sensor CMOS ISO (REF 100 ISO)-PSNR
equipped cameras, firstly images are captured with CMOS and CCD
PSNR ( with ref to 100 ISO)
sensor based cameras. Then the analysis of images is done in
MATLAB and MSE, PSNR and SSIM are evaluated for images
captured by CMOS (Targeted Image) and CCD (Reference Images) 95
sensor based cameras.
CCD VS CMOS-MSE 90
MSE 85
0.00006 80
A B C D
0.00004 b)
CMOS ISO ( WITH REF TO 100
0.00002 ISO)-SSIM
SSIM ( with ref to 100 ISO)
0
0.55
-2 -1 0 1 2

a)
CCD VS CMOS-PSNR 0.5
PSNR
150 0.45
A B C D
100 c)
Fig.8: Variation in MSE, PSNR and SSIM values on different
50
Exposure Values for CMOS Images. A, B, C, D represents the ISO
levels between 100-200, 100-400, 100-800, 100-1600 respectively
CMOS SHUTTER SPEED
0 (REF 1/50)-MSE
-2 -1 0 1 2
0.00006
b)
0.00004
CCD vs CMOS-SSIM

SSIM 0.00002
0.3
0
0.2 0.02 0.033 0.04 0.05
a)
0.1
CMOS SHUTTER SPEED (REF 1/50)-
PSNR
0
-2 -1 0 1 2 0.54
c) 0.52
Fig.7: Variation in MSE, PSNR and SSIM values on different
0.5
Exposure Values for CMOS & CCD Images
It has been observed that MSE between two images is of 0.000042, 0.48
which means there is hardly difference between two images and that 0.02 0.033 0.04 0.05
of PSNR is 91.95, which depict that the images are of good quality.
SSIM for both images id found to very low. Images were also b)
captured and different ISO levels and I is found that the error in the
images are increased as ISO levels are increased. As a result, PSNR

IJARI 17

Electronic copy available at: https://ssrn.com/abstract=3559236

 International Conference of Advance Research and Innovation (ICARI-2020)
Communications and Signal Processing (WCSP), pp 01-05, Oct
CMOS shutter speed (Ref 1/50)-SSIM
2014.
[11] Erwin, Nevriyanto A., Purnamasari D., “Image Enhancement
0.54
Using the Image Sharpening, ContrastEnhancement, and
0.53 Standard Median Filter (Noise Removal) with Pixel-Based and
Human Visual System-Based Measurements”, International
0.52
Conference on Electrical Engineering and Computer
0.51 Science(ICECOS), pp 114-119, Aug 2017.
[12] Hore A., Ziou D., “Image quality metrics: PSNR vs. SSIM”, 20th
0.5
International Conference on Pattern Recognition, pp 2366-2369,
0.49 Aug 2010.
[13] Tan H.L., Li Z., Tan Y. H., Rahardja S., Yeo C., “A Perceptually
0.48 relevant MSE-Based Image Quality Metric”, IEEE Transactions
0.02 0.033 0.04 0.05 on Image Processing ,Volume: 22 , Issue: 11,pp 4447- 4459,
c) July 2013.
Fig. 9: Variation in MSE, PSNR and SSIM values on different [14] Hirayama T., “ The Evolution of CMOS image sensos”, IEEE
Shutter Speed Asian Solid-State Circuits Conference (A-SSCC), pp 05-08,
Fig. 9 depicts the variation in MSE, PSNR and SSIM values on Nov 2013.
different Shutter Speed. The SSIM also kept on deceasing due to [15] Joshi K., Yadav R., Allwadhi A., “ PSNR and MSE Based
noise as the ISO levels increases. Here, one more factor which Investigation of LSB”, International Conference on
consider in photography is considered, I.e. shutter speed. It had been Computational Techniques in Information and Communication
observed that the value of MSE increases as the shutter speed is Technologies (ICCTICT), pp 01-05, July 2016.
reduced and correspondingly value of PSNR is also reduced. SSIM [16] Wang J., Liu Y., Wei P., Tian Z., Li Y., Zheng N., “Fractal
also decreases with reduced shutter speed. The results of MSE, PSNR Image Coding Using SSIM “, 18thIEEE International
and SSIM for ISO and shutter speed of the camera. Conference on Image Processing, pp 241-244, Dec 2011.
6. Conclusions [17] Xue F., Ye Z., Lu., Liu H., Li B., “MSE period based estimation
This paper tries to bridge up the gap between engineering and of first quantization step in double compressed JPEG images”,
photographic aspects of the images. One can observe for what value Signal Processing: Image Communication, volume 57, pp 76-
of ISO; Exposure Value or Shutter Speed can effect on the quality of 83, may 2017.
an image. Further, it has been noticed that quality of CCD image
sensor is still better than CMOS but due to numerus reason CMOS
image sensor are better. Considering this fact, again technical and
photographic parameters are analyzed and results found to be
satisfactory in comparison to CCD Image sensors. Hence, the
objective of this paper are achieved with help of graph and
mathematical values.
References
[1] Jerram P., Stefanov K. CMOS and CCD image sensors for space
applications, High Performance Silicon Imaging (Second
Edition), Fundamentals and Applications of CMOS and CCD
Sensors, Woodhead Publishing Series in Electronic and Optical
Materials 2020, pp 255-287, oct 2019.
[2] Centen P. Complementary metal-oxide-semiconductor (CMOS)
and charge coupled device (CCD) image sensors in high-
definition TV imaging, High Performance Silicon Imaging
(Second Edition), Fundamentals and Applications of CMOS and
CCD Sensors, Woodhead Publishing Series in Electronic and
Optical Materials 2020, pp 437-471, oct 2019.
[3] Lustica A.CCD and cCMOS Image sensors in New HD cameras,
Proceedings ELMAR-2011, pp 133-136, Oct 2011.
[4] Ang, T.(2003). Advanced Digital Photography. London:
Mitchell Beazley.
[5] Ang T. (2003). Digital Photography An Introduction. London:
Dorling Kindersley.
[6] Mehta S., Patel A., Mehta J.CCD or CMOS image sensor for
photography, International Conference on Communications and
Signal Processing (ICCSP), pp 0291-0294, Nov 2015.
[7] Prasantha H.S., Shashidhara H.L., Balasubramanya K.N., “
Image Scaling Camparison using universal image quality index”
, International Conference on Advances in Computing, Control,
and Telecommunication Technologies, pp 859-863, dec 2009.
[8] Centen P., “Complementary metal-oxide-
semiconductor(CMOS) and charge coupled device (CCD)
image sensors in high-definition TV imaging”, High
Performance Silicon Imaging , Fundamentals and Applications
of CMOS and CCD Sensors 2014, pp 373-410, may 2014.
[9] Bigas M., Cabruja E., Forest J., Salvi J., “Review of CMOS
image sensors” Volume 37, Issue 5, pp 433-45, Sept 2005.
[10] Cui Z., Gan Z., Tang G., Liu F., Zhu X., “Simple and effective
image quality assessment based on edge enhanced Mean square
Error”, Sixth International Conference on Wireless

IJARI 18

Electronic copy available at: https://ssrn.com/abstract=3559236