2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science

An Analysis of Image Enhancement

2024 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS) | 979-8-3503-4846-0/24/$31.00 ©2024 IEEE | DOI: 10.1109/SCEECS61402.2024.10482109

Techniques
Reena Sharma 1,2 Nikita Jain Sudhanshu Tiwari
1 Dept. of Computer& Comm. Engineering Dept. of Computer Engineering Dept. of Advance Computing
Manipal University Poornima College of Engineering Poornima College of Engineering
Jaipur, India Jaipur, India Jaipur, India
2 Dept. of Advance Computing [email protected] [email protected]
Poornima College of Engineering
Jaipur, India
[email protected]

Astha Gothi
Dept. of Advance Computing
Poornima College of Engineering
Jaipur, India
[email protected]

Abstract – A crucial aspect of image processing is image low resolution picture. For certain specialized uses,
enhancement. Image enhancement techniques can be used the outcome is better suited than the original image.
to improve the image's quality and aesthetic appeal in
There are many different kinds of filtering
order to more accurately convey a transform. In general,
multiple types of photos, including satellite images, techniques that we may apply to improve the image
medical images, and photographs taken in real life, are quality. A single, specialized image enhancement
impacted by low contrast and unique noise. A picture can filter may not give the optimum quality throughout a
be enhanced or modified using the filtering process. For complicated image in many situations. In order to
instance, we may filter an image to highlight some
eliminate Gaussian noise from smooth areas of a
elements while removing others. Filtering is used to
accomplish image processing operations including edge picture, for instance, a filter would often seek to blur
enhancement, sharpening, and smoothing. Image edge detail. Alternatively, a non-linear filter meant to
enhancement, filtering, segmentation etc. all are maintain edge detail could result in unfavorable
considered in image preprocessing. artifacts in smooth, noisy parts of a picture.
Keywords- Image enhancement, modifying, filtering, edge In computer vision, each linear or nonlinear filter
enhancement. would be applied to the areas of the image where it
would produce the greatest results in order to
I. INTRODUCTION produce the best possible outcome. This strategy
calls for a mechanism, nevertheless, to choose the
Image processing used to highlight a certain element of optimum filter from a range of options for each point
an image is known as image enhancement [1]. Image in the image. Image enhancement techniques are also
enhancement fundamentally involves making images used in medical field for diagnose the disease [6].
easier for humans to understand or interpret while also The purpose of image enhancement is to make
giving other automated image processing methods for pictures easier for humans to understand or perceive
better input. The main goal of image enhancement is to or to provide other automated image processing
change image’s characteristics so that they are more methods for better result [5].Two general categories
suited to a certain purpose and viewer. may be used to classify image enhancing techniques:
One or more image characteristics are changed 1. Pixel-based approaches that act in the spatial
throughout this operation. A provided job determines domain.
the qualities to be used and how they are updated. The
choice of image augmentation techniques will also be 2. Frequency based methods, which give the results
very subjective due to observer-specific elements, for high and low both frequency with an image's
including the human visual system and the observer's Fourier transform.
experience. The following scenarios involve the use of
The most basic yet necessary image processing
image enhancement: removing the image's noise, the
processes are the pixel processing techniques, which
highlighting of object edges and improvement of the

979-8-3503-4846-0/24/$31.00 ©2024 IEEE

Authorized licensed use limited to: HKBK College of Engineering. Downloaded on February 13,2025 at 03:49:33 UTC from IEEE Xplore. Restrictions apply.
 are mostly employed for contrast amplification [7]. In Equalization of the Global Histogram, the global
Image Negative has uses in medical imaging and is information of the image is the foundation for global
suitable for highlighting white detail embedded in dark histogram equalization (GHE). First, a window is
regions. In general, contrast manipulation can benefit taken into consideration for the input picture I in
from power-law transformations. For spatial domain or Local Histogram Equalization (LHE) under Block
frequency domain, we use linear smoothing filter for Overlapped Histogram Equalization [3].
In Local Histogram Equalization, we start with the
remove the noise [8].
first pixel of the image, traversing begins with the
For a dark image, a power-law transformation with a window's center point. Equalize the center point and
fractional exponent is used to expand the range of grey move the window to the neighboring pixel to
levels. The higher-level values of the log perform equalization. This process is performed for
each pixel in the input picture, pixel by pixel. The
transformation are useful for increasing features in the
input image is transformed into the high contrast
darker areas of the image at the price of detail in the
image after LHE. But LHE is also not the best
brighter areas. A power-law transformation with higher technique since the image is more -enhanced and this
than 1 is used to produce a compression of grey levels process becomes more difficult as the image size
for a picture that has a washed-out appearance in image grows. In LHE, different transformation of the same
processing [9]. By the help of enhanced image, we get gray scale at different place is perform for enhance
the required details very easily and accurately in the images.
comparison to noisy images. The main objective of
image enhancement is to provide accurate details for Many authors proposed different techniques of
further processing in image processing. Histogram Equalization. Illumination problem is
solved by proposed new histogram technique on
facial images [10]. In Quadrant Dynamic Histogram
Equalization, Histogram is divided into four
II. HISTOGRAM EQUALIZATION TECHNIQUES quadrants on the base of the original image’s median
value. [11]. In Bi – Histogram Equalization
technique, processing time is decrease in
Histogram Equilazation Techniques are very simple comparison to other techniques [12]. In low dynamic
and easy approach for image enhancement. In all of range images, Genetic Algorithm is applied for
Histogram technique, some common techniques we image enhancement. [13]. Many algorithm proposed
implement here and compare these techniques. for image enhancement using histogram equalization
Globalization of Histograms when a low-contrast and they all have some advantages and
picture with grey levels ranging from 0 to L-1 is used disadvantages.
as the input, the likelihood of a pixel having the
intensity value k is given by
III. EXPERIMENTAL RESULTS AND
Pk = nk/N (1) COMPARISON

Where N is the total number of pixels in a picture, k is For image enhancement when we apply Global
the intensity value ranging from 0 to L-1, and nk is the histogram technique and local histogram technique
number of instances of the kth intensity in the image. on an image then we get resultant image as shown in
In Histogram Equalization, by the distribution of “Fig 3” and “Fig 5”. Resultant histogram of Global
intensity, contrast of image is adjusted. This is histogram technique and Local histogram technique
basically related to the linear cumulative probability are shown in “Fig 4 and “Fig 6” respectively.
function which is associated to the images. Formula for
expressing the cumulative distribution function (CDF)
in terms of the probability distribution function

CDF = P1+P2……..PL-1 (2)

The problem with GHE is that it high-enhances pixels

with high probability while less-enhances pixels with
low probability, which might decrease the quality of
the target item in the picture [2]. Global Histogram
Equilazation enhance the whole image.
Fig 1: Original Image

SCEECS 2024
Authorized licensed use limited to: HKBK College of Engineering. Downloaded on February 13,2025 at 03:49:33 UTC from IEEE Xplore. Restrictions apply.
 Fig 2: Histogram of Original Image
Fig 5: Image after apply LHE technique on original image

Fig 3: Image after apply the GHE technique Fig 6 : Histogram after apply the LHE on original image

We can compare all images by calculate Peak Signal to

Noise Ratio (PSNR). PSNR ration is used for measure the
quality of reference image and output image. As we
know, when PSNR value is higher than it is better image.
Comaprison table of PSNR of original image, apply LHE
on original image, apply GHE on original image show in
table 1.

Table 1: PSNR values of images

Image PSNR value

Original image 23.67

Image after applied GHE 22.87

Image after applied LHE 32.89

Fig 4 : Histogram of updated image ( Fig 3)

SCEECS 2024
Authorized licensed use limited to: HKBK College of Engineering. Downloaded on February 13,2025 at 03:49:33 UTC from IEEE Xplore. Restrictions apply.
 IV. CONCLUSION detection using chest X-ray images”, Computers
in Biology and Medicine, Volume 132, 104319,
A broad range of methods for changing images to ISSN 0010-4825, 2021.
produce aesthetically appealing images are available
[7] Dibya Jyoti Bora, “Importance of Image
through image enhancement algorithms. The exact task, Enhancement Techniques in Color Image
picture content, observer characteristics, and viewing Segmentation: A Comprehensive and Comparative
circumstances all play a role in the technique selection. Study”, Indian J.Sci.Res. 15 (1): 115-131, 2017.

An image's histogram, which is a depiction of the gray [8] Distante, A., Distante, C., “Image Enhancement
scale frequencies, contains crucial details about an Techniques”. In: Handbook of Image Processing and
image's contrast. A technique called histogram Computer Vision. Springer, Cham, 2020.
equalization increases contrast by equally spreading the
[9] Hussain, K., Rahman, S., Rahman, M.M. et al. A
values of the grey levels. The only fully automated
histogram specification technique for dark image
process that can be used is the global histogram enhancement using a local transformation method.
equalization. Although it was not covered in this article, IPSJ T Comput Vis Appl 10, 3
the computational cost of enhancement algorithms may https://doi.org/10.1186/s41074-018-0040-0, 2018.
be a key factor in selecting an approach for real-time
applications. [10] A. Salah-eldin, K. Nagaty, and T. Elarif, “An
Enhanced Histogram Matching Approach Using the
Despite the fact that each of these algorithms works well Retinal Filter ’ s Compression Function for
when used alone, in reality, one has to combine many of Illumination Normalization in Face Recognition,” in
International Conference on Image Analysis and
these techniques to improve images.
Recognition, 2008, pp. 873–883.

ACKNOWLEDGEMENT [11] C. H. Ooi and N. A. M. Isa, “Quadrants dynamic

histogram equalization for contrast
The author would like to thank Poornima College of enhancement,” IEEE Trans. Consum. Electron.,
Engineering for provide the research reported in this vol. 56, no. 4, pp. 2552–2559, 2010.
paper.
[12] C. H. Ooi, N. P. Kong, and H. Ibrahim, “Bi-
REFERENCES histogram equalization with a plateau limit for
digital image enhancement,” IEEE Trans. Consum.
Electron., vol. 55, no. 4, pp. 2072–2080, 2010.
[1] A.Rosenfeld and A.C.K ak, “Digital Picture
processing”, computer science and Applied [13] S. Hashemi, S. Kiani, N. Noroozi, and M. E.
mathematics, Academic Press, 1982. Moghaddam, “An image contrast enhancement
method based on genetic algorithm,” Pattern
[2] Cheng Lei, and Lv Weijie, “A Fast Algorithm for foggy Recognit. Lett., vol. 31, no. 13, pp. 1816–1824,
image contrast enhancement” International conference on 2010.
TMEE, 2011.

[3] Lan GE Xin-Yu Zhangan and Tian-Fu Wang,

“Entropy based Local Histogram Equalization for
Medical Ultrasound Image Enhancement”, IEEE
2008.

[4] Raman Maini and Himanshu Aggarwal , “A

Comprehensive Review of Image Enhancement
Techniques” , Journal of computing, volume 2, issue
3, ISSN 2151-9617, March 2010.

[5] Narayan, L.K., Vishwakarma, V.P. “A Review on

Different Image Enhancement Techniques”. In: Peng, SL.,
Jhanjhi, N.Z., Pal, S., Amsaad, F. (eds) Proceedings of 3rd
International Conference on Mathematical Modeling and
Computational Science. ICMMCS 2023. Advances in
Intelligent Systems and Computing, vol 1450. Springer,
Singapore, 2023.

[6] Tawsifur Rahman, Amith Khandakar, Yazan

Qiblawey, Anas Tahir, Serkan Kiranyaz, Saad Bin
Abul Kashem, Mohammad Tariqul Islam, Somaya Al
Maadeed, Susu M. Zughaier, Muhammad Salman
Khan, Muhammad E.H. Chowdhury, ‘Exploring the
effect of image enhancement techniques on COVID- 19

SCEECS 2024
Authorized licensed use limited to: HKBK College of Engineering. Downloaded on February 13,2025 at 03:49:33 UTC from IEEE Xplore. Restrictions apply.