EDGE PRESERVED IMAGE DENOISING

BY USING DIRECTIONAL DIFFERENCE

BASED MEDIAN FILTER
PROJECT REPORT
PHASE I
Submitted by

ARUN.M

(080107127013)

JAGANNATHAN.A.J.S

(080107127040)

KAJAHUSSAIN.A

(080107127043)

AVINASH VIKRAM.M

(090407127001)

in partial fulfillment for the award of the degree

of

BACHELOR OF ENGINEERING
in

ELECTRONICS AND COMMUNICATION ENGINEERING

TAMILNADU COLLEGE OF ENGINEERING, COIMBATORE
ANNA UNIVERSITY OF TECHNOLOGY COIMBATORE,
COIMBATORE-641047
OCTOBER 2011

ANNA UNIVERSITY OF TECHNOLOGY COIMBATORE,

COIMBATORE-641047

BONAFIDE
CERTIFICATE

Certified that this project report EDGE PRESERVED IMAGE DENOISING BY

USING DIRECTIONAL DIFFERENCE BASED MEDIAN FILTER is the
bonafide work of ARUN.M, JAGANNATHAN.A.J.S, KAJAHUSSAIN.A and
AVINASH VIKRAM.M who carried out the project work under my supervision.

SIGNATURE

SIGNATURE

Dr.M.KARTHIKEYAN, Ph.D.,

Mrs.G.SANTHANAMARI, M.E., (Ph.D)

HEAD OF THE DEPARTMENT

SUPERVISOR
ASSISTANT PROFESSOR

Department of ECE,

Department of ECE,

Tamilnadu College of engineering,

Tamilnadu College of engineering,

Coimbatore-641659.

Coimbatore-641659.

Submitted for the Anna University Examination held on ..

INTERNAL EXAMINER

EXTERNAL EXAMINER

ABSTRACT
Digital images are often corrupted by many types of noise
including salt-and-pepper, which are normally affect the acquisition and
transmission. Impulse noise which is a set of isolated pixels can make a great
difference on the configuration of image. It is often caused by malfunctioning
pixels in camera sensors, faulty memory locations in hardware or transmission
of the image in a noisy channel. It is essential to eliminate salt-and-pepper noise
in the image and to preserve the image edge and integrity. The proposed method
consists of a detection stage where the noisy pixels are detected and followed by
filtering which replaces only the noisy pixels. The noisy pixels are detected by
taking the minimum and maximum gray level intensity values in the sliding
window. Then these detected noisy pixels will be restored by the median value
of the noise free pixels in the sliding window, and then the directional
difference median filtering is used to handle the conflict of noise suppression
and edge-preserving. The proposed method is convenient and efficient on
removing salt and pepper noise in the image and at the same time can achieve
good performance of edge preserving, even when the image is corrupted by salt
and pepper noise of higher density. Simulation results shows that it can remove
salt and pepper noise effectively with edge-preserving compared to other
existing nonlinear filters. Thus the proposed method is simple to realize and it is
simulated by using MATLAB.

ii

TABLE OF CONTENTS
CHAPTER NO.

TITLE

PAGE NO.

ABSTRACT

ii

LIST OF FIGURES

vi

LIST OF TABLES

vii

LIST OF ABBREVIATIONS

viii

INTRODUCTION

1.1 Need For The Project

1.2 Motivation

1.3 Original Contributions Of The Thesis

1.4 Organization Of The Project

SYSTEM ANALYSIS

2.1 Existing System

2.1.1. Drawbacks

2.2 Proposed System

2.3 Feasibility Study

2.3.1 Economical Feasibility

2.3.2 Operational Feasibility

2.3.3 Technical Feasibility

iii

SYSTEM SPECIFICATION

3.1.Hardware Requirements

3.2.Software Requirements

SOFTWARE DESCRIPTION

10

4.1 Matlab Description

10

4.2 Features

11

4.3 Image Processing Tool Box

12

4.3.1 Features Of Image Processing

12

Tool box

4.3.1.1Pre-processing and Post-processing Images

13

4.3.1.2 Image Analysis

13

4.3.1.3 Image Enhancement

14

4.3.1.4 Image Deblurring

15

PROJECT DESCRIPTION

17

5.1 Problem Definition

17

5.2 Overview Of The Project

17

5.3 Description Of The Algorithm

18

5.3.1 Noise Detection Algorithm Along with

Adaptive Median Filtering

iv

18

5.3.2 Directional Difference Filtering Algorithm

19

5.4 Data Flow Diagram

21

5.5 Data Base Design

22

RESULTS AND DISCUSSION

24

6.1 Simulation Results

24

6.2 Performance Analysis

25

6.2.1comparison Table For 512x512 Lena Image

25

6.2.2Comparison Table for Peppers Image

27

CONCLUSION

31

7.1 Future Enhancements

31

REFERENCES

32

LIST OF FIGURES
FIGURE

TITLE

NO.

PAGE NO.

4.1

Performing Connected Components Analysis

15

4.2

Image Of The Sun Using Deblurring Algorithms

16

5.1

Fundamental Steps In Digital Image Processing

21

5.2

Standard Testing Images Used In Existing Algorithms

22

5.3

Standard Testing Images Used In Proposed Algorithm

23

6.1

Simulation Results For 60% Corrupted Lena Image

26

6.2

Simulation Results For 60% Corrupted 'Peppers Image

28

6.3

Simulation Results For 60% Corrupted 'Cameraman

6.4

Image

29

Simulation Results For 60% Corrupted 'Rahman Image

30

vi

LIST OF TABLES

TABLE NO.

TITLE

PAGE NO.

Comparison Of PSNR Values Of Proposed Technique h

6.1

With Existing Technique For 512x512 Lena Image.

25

Comparison Of MSE Values Of Proposed Technique

6.2

With Existing Technique For 512X512 LENA IMAGE

25

Comparison Of PSNR Values Of Proposed Technique

6.3

With Existing Technique For 384x512 Peppers Image

27

Comparison Of MSE Values Of Proposed Technique

6.4

With Existing Technique For 384x512 Lena Image

vii

27

LIST OF ABBREVIATIONS
ACWM

Adaptive impulse detector with Center- Weighted Mean

ATMA

Alpha-Trimmed Mean-based Approach

BDND

Boundary Discriminative Noise Detection

BLAS

Basic Linear Algebra Subprograms

CWM

Center Weighted median Filter

DBA

Decision Based Algorithm

DCT

Discrete Cosine Transform

DICOM

Digital Imaging & Communication in Medicine

DWMF

Directional Weighted Median Filter

EDRIN

Efficient Denoising Chip For The Removal Of Impulse

Noise

EEPA

Efficient Edge-Preserving Approach

EISPACK

Eigen vector Solution Package

FFT

Fast Fourier Transform

GUI

Graphical User Interface

ICC

International Color Consortium

LINPACK

Linear algebra Package

MSE

Mean Square Error

OCS

Open Close Sequence

PSNR

Peak Signal to Noise Ratio

RDRIN

Robust Detection technique for removing

Random valued Impulse Noise

SAWM

Switching based Adapted Weighted Mean

SMF

Switching Median Filter

viii

CHAPTER 1

INTRODUCTION
1.1 NEED FOR THE PROJECT
Image sequences have conquered their place among the most important
information carriers in today's world. Their applications such as broadcasting,
video-phone, traffic observations, surveillance systems, autonomous navigation
and so on. The field of digital image processing refers to processing digital
image by means of a digital computer. It encompasses processes that extract
attributes from images including the recognition of individual objects.
Digital images are often corrupted by many types of noise including
salt-and-pepper noise, which are normally acquired during image acquisition
and transmission. Impulse noise which is a set of isolated pixels can make a
great difference on the configuration of image. It is essential to eliminate saltand-pepper noise in the image and preserve the image edge and integrity.
Normally the pixels which have maximum and minimum gray level intensity
value in the sliding window are the pixels that are corrupted by the salt and
pepper noise.
The proposed method concentrates on image denoising where impulse
noises in images are removed. Impulse noise is found in situations where quick
transients such as faulty switching, during imaging, due to transmission errors,
malfunctioning pixel elements in the camera sensors, faulty memory locations,
and timing errors in analog-to-digital conversion. An important characteristic of
this type of noise is that only part of the pixels is corrupted and the rest are
noise-free.
This paper deals with the image restoration which is a highly
developing and gaining importance, because of the significant increase in the
use of digital images over the internet and in enhancement of the images in
various fields of research like.
1

 Medical applications and Defence applications

Science and engineering fields.
The image processing is considered to be very important in these for
the pre-processing and the post-processing processes.

1.2 MOTIVATION
The acquisition or transmission of digital images through sensors or
communication channels is often interfered by impulse noise. Noise removal
from a corrupted image is finding vital application in image transmission over
the wideband network. It is imperative and even indispensable, to remove these
corrupted pixels to facilitate subsequent image processing operations, such as
edge detection, image segmentation and object recognition.
In digital image processing, noise reduction is one of basic preprocessing steps usually include image enhancement methods. The aim of
impulse noise reduction is to suppress the noise while preserving the important
fine details and edges which are the two different boundary regions. The
performance of some tasks in next levels such as segmentation/classification for
various applications such that biometric recognition depends on the success of
noise reduction in previous level.
The purpose of image restoration is to "compensate for" or "undo"
defects which degrade an image. Degradation comes in many forms such as
motion blur, noise and camera misfocus. In cases like motion blur, it is possible
to come up with a very good estimate of the actual blurring function and "undo"
the blur to restore the original image. In cases where the image is corrupted by
noise, the best we may hope to do is to compensate for the degradation it
caused. In this project, a new algorithm for effective restoration of degraded
images is proposed.

1.3 ORIGINAL CONTRIBUTIONS OF THE THESIS

The main parameters to be considered during restoration of a distorted
image are as follows:
Miss detection rate: The rate at which the noisy pixels in the corrupted
image are missed during the process of detection of noisy pixels.
False alarm rate: The rate at which the noise free pixels are indicated
wrongly as noisy pixels.
Peak Signal to Noise Ratio (PSNR): The phrase Peak Signal-to-Noise
Ratio, often abbreviated PSNR, is an engineering term for the ratio
between the numbers of noise free pixels to the number of noisy pixels of
a corrupted image.
PSNR=

Noise density: The amount of noise added to the original image.

Processing time: The time required for the complete execution of the
program.
MSE: Mean Square Error is defined as the average of the square of the
difference between the desired response and the actual system output(the
error)

MAE: Mean Absolute Error is a quantity used to measure how closed

forecasts or predictions are to the eventual outcomes.
MAE=

 SSIM: Structural SIMilarity (SIMM) index is a method for measuring the

similarity between two images.
The contributions of this thesis can be generalized as follows:

The proposed algorithm has high ability in identifying impulse noise

thereby achieving zero miss detection rate and false alarm rate.

This algorithm provides high PSNR and low MSE values.

The processing time for this algorithm is very less.

1.4 ORIGANIZATION OF THE PROJECT

The thesis is composed of eight chapters. The structure is as follows:

In Chapter 1, the introduction and need for the project are given.

Chapter 2 describes existing systems with its drawbacks and the proposed

system are analysed with its feasibility study.

Chapter 3 describes the system specifications for both hardware and

software requirements.

Chapter 4 describes the MATLAB software in basic applications of

image processing and numerical calculations along with the Image Processing
Toolbar and its features.

Chapter 5 deals with the description, problem definition and overview of

the project.

In Chapter 6, simulation results and the performance analysis

comparisons between the existing and proposed algorithm are discussed.

Chapter 7 describes the conclusion and future enhancement of the project

is discussed.

In Chapter 8, references are given

CHAPTER 2

SYSTEM ANALYSIS
2.1 EXISTING SYSTEM
The median filter is the basic nonlinear filter for removing impulse
noise. It has good noise suppression ability and high computational efficiency
but it is prone to damage such important details as thin lines and sharp corners
since it replaces every pixel by the median value of its neighbouring pixels.
The recent methods which are effective in the removal of random valued
impulse noise are a new Directional Weighted Median Filter (DWMF) for
removal of random valued impulse noise [2], an Efficient Denoising chip for the
Removal of Impulse Noise (EDRIN) [14] And Robust Detection technique for
removing Random valued Impulse Noise (RDRIN) [10]. The DWMF uses
directional weighted filters which are effective in removing very low noise
densities and is not efficient for higher noise densities.
.2.1.1 Drawbacks
The principle drawback is that they have limited performance in terms
of false alarm and miss detections. Hence, they cannot preserve the image
details and edges, especially when the noise is high.
Median filter and other versions of median filters have a good ability,
in the low noise ratios and on the other hands in case of the high noise ratio,
these methods causes blurring image and actually destroys vital textures. In
these cases, the relationship between noise reduction and preservation of
important details of image is very important and needs proper algorithms for
noise reduction. The major drawback of the conventional vector median
approaches is that they apply median operation to each pixel, irrespective of it
being corrupted or not.

This approach is quite successful in handling the signal dependent

noise. Impulse noise can also be removed using higher order statistics. But this
method involves computation of higher order statistical terms, which are
computationally expensive [2]. The EDPA, EEPA and ACWN methods only
perform well when an image is corrupted with 50% of salt and pepper noise or
lower [8].
2.2 PROPOSED SYSTEM
In the proposed method there is one stage of detection and two stage of
filtering. In the first stage, the maximum and minimum gray level intensities in
the sliding window of the noisy image are considered as the threshold value to
detect the presence of noise. In first stage of filtering, these detected noisy
pixels are replaced by the median value of the noise free pixels. In second stage
of filtering directional difference median filtering is used for noise suppression
and edge-preservation. The proposed method is convenient and efficient on
removing salt and pepper noise in the image, and at the same time can achieve
good performance of edge preserving, even when the image is corrupted by salt
and pepper noise of high density.
2.3 FEASIBILITY STUDY
A feasibility study is an evaluation of a proposal designed to determine
the difficulty in carrying out a designated task. Generally, a feasibility study
precedes technical development and project implementation. In other words, a
feasibility study is an evaluation or analysis of the potential impact of a
proposed project.
Feasibility analysis carried out for this project is as follows,
2.3.1 Economical feasibility
For any system if the expected benefits equal or exceed the expected
costs, the system can be judged to be economically feasible. In economic

feasibility, cost benefit analysis is done in which expected costs and benefits are
evaluated. Economic analysis is used for evaluating the effectiveness of the
proposed system. In economic feasibility, the most important is cost benefit
analysis. An impulse noise detection algorithm for switching median filters is
computationally expensive, but it is simulated and seems to be an exceedingly
effective and accurate algorithm for impulse noise detection methods and this
proposed scheme provides better image restoration comparison with the other
existing conventional methods.
2.3.2 Operational feasibility
Operational feasibility is a measure of how well a proposed system
solves the problems and takes advantage of the opportunities identified during
scope definition and how it satisfies the requirements identified in the
requirements analysis phase of system development.
The highly effective directional median filter behaviour is simpler and
it is easy to implement and preserve image details by selecting only noise
pixels for processing and will speed up the filtering process. They can achieve
good edge preserving performance employing fuzzy sets. With the directional
difference median filter it is useful and effective to removal salt and pepper
noise in image processing compared to the conventional median based filters.
The proposed method provides amazing accurate results even when noise
density is as high as 60%.
2.3.3 Technical feasibility
This project is developed using MATLAB. It is executed in the
minimum hardware of Intel Pentium III,RAM of 1 GB, hard disk capacity of
80GB,15 inches monitor,104 keys and mouse, the software MATLAB, the
windows XP or Linux operating systems and that has been used in this project
are found to be technically feasible.
8

CHAPTER 3

SYSTEM SPECIFICATION

3.1 HARDWARE REQUIREMENTS

PROCESSOR

Pentium III (or higher)

SPEED

2.53 GHz (or higher)

RAM

1 GB (or higher)

HARD DISK DRIVE

40 GB (or higher)

MONITOR

15 Color (or higher)

KEYBOARD

Logitech 104 Keys (or

Higher)

3.2 SOFTWARE REQUIREMENTS

OPERATING SYSTEM

Windows xp3 (or higher) or

Linux (Red hat or Ubuntu)

DEVELOPMENT TOOL

MATLAB 7.0 (or higher)

CHAPTER 4

SOFTWARE DESCRIPTION
4.1 MAT LAB DESCRIPTION
MATLAB is a high-level technical computing language and interactive
environment for algorithm development, data visualization, data analysis and
numeric computation. Using the MATLAB product, we can solve technical
computing problems faster than with traditional programming languages, such
as C, C++ and FORTRAN. MATLAB is an integrated technical computing
environment that combines numeric computation, advanced graphics and
visualization and a high-level programming language.
The MATLAB can be used in a wide range of applications, including
signal and image processing, communications, control design, test and
measurement, financial modeling and analysis and computational biology.
Add-on toolboxes (collections of special-purpose MATLAB functions,
available separately) extend the MATLAB environment to solve particular
classes of problems in these application areas.
MATLAB provides a number of features for documenting and sharing
our work. We can integrate our MATLAB code with other languages and
applications and distribute our MATLAB algorithms and applications.
MATLAB is an interactive system whose basic data element is an array
that does not require dimensioning. It allows us to solve many technical
computing problems, especially those with matrix and vector formulations, in a
fraction of the time it would take to write a program in a scalar non-interactive
language such as C or FORTRAN.
The name MATLAB stands for matrix laboratory. MATLAB was
originally written to provide easy access to matrix software developed by the
10

LINPACK and EISPACK projects. Today, MATLAB engines incorporate the

LAPACK and BLAS libraries, embedding the state of the art in software for
matrix computation.
The MATLAB language is a high-level matrix/array language with
control flow statements, functions, and data structures, input/output and objectoriented programming features. It allows both "programming in the small" to
rapidly create quick programs we do not intend to reuse. We can also do
"programming in the large" to create complex application programs intended for
reuse.
4.2 FEATURES
The various features of MATLAB are as following
Development environment for managing code, files and data.
Interactive tools for iterative exploration, design and problem solving.
Mathematical functions for linear algebra, statistics, Fourier analysis,
filtering, optimization and numerical integration.
2-D and 3-D graphics functions for visualizing data.
Tools for building custom graphical user interfaces.
Functions for integrating MATLAB based algorithms with external
applications and languages, such as C, C++, FORTRAN, Java, COM
and Microsoft Excel.
Packages MATLAB applications as executable and shared libraries.
Allows

distribution

components

of

standalone

executable

and

software

royalty-free.

Allows incorporation MATLAB based algorithms into applications

developed using other languages and technologies.
Encrypts MATLAB code so that it cannot be viewed or modified.
11

4.3 IMAGE PROCESSING TOOLBOX

The image processing tool box is normally used to perform image
processing, analysis and algorithm development.

Image Processing Toolbox

software provides a comprehensive set of reference-standard algorithms and

graphical tools for image processing, analysis, visualization and algorithm
development. We can restore noisy or degraded images, enhance images for
improved intelligibility, extract features, analyze shapes and textures and
register two images. Most toolbox functions are written in the open MATLAB
language, giving us the ability to inspect the algorithms, modify the source code
and create our own custom functions.
We can perform image enhancement, image deblurring, feature
detection, noise reduction, image segmentation, spatial transformations and
image registration. Many functions in the toolbox are multithreaded to take
advantage of multicore and multiprocessor computers. Image Processing
Toolbox supports a diverse set of image types, including high dynamic range,
giga pixel resolution, ICC-compliant colour and tomography images. Graphical
tools let we explore an image, examine a region of pixels, adjust the contrast,
create contours or histograms and manipulate regions of interest (ROIs). With
the toolbox algorithms we can restore degraded images, detect and measure
features, analyze shapes and Textures and adjust the colour balance of images.
4.3.1 FEATURES OF IMAGE PROCESSING TOOLBOX

Image enhancement, filtering and deblurring.

Image analysis, including segmentation, morphology, feature extraction

and measurement.

Spatial transformations and image registration.

Image transforms, including FFT, DCT, Radon and fan-beam projection.

12

Workflows for processing, displaying and navigating arbitrarily large

images.

Modular interactive tools, including ROI selections, histograms and

distance measurements.

ICC colour management.

Multidimensional image processing.

Image-sequence and video display.

DICOM import and export.

4.3.1.1 Pre-processing and Post-processing images

Image Processing Toolbox provides reference-standard algorithms for
pre-processing and post processing tasks that solve frequent system problems,
such as interfering noise, low dynamic range, out-of-focus optics and the
difference in colour representation between input and output devices.
4.3.1.2 Image analysis
Image Processing Toolbox provides a comprehensive suite of reference
standard algorithms and graphical tools for image analysis tasks such as
statistical analysis, feature extraction and property measurement.
Statistical functions let you analyze the general characteristics of an
image by:

Computing the mean or standard deviation.

Determining the intensity values along a line segment.

Displaying an image histogram.

Plotting a profile of intensity values.

Edge-detection algorithms let you identify object boundaries in an

image. These algorithms include the Sobel, Prewitt, Roberts, Canny and

13

Laplacian of Gaussian methods. The powerful canny method can detect true
weak edges without being "fooled" by noise.
Image segmentation algorithms determine region boundaries in an
image. We can explore many different approaches to image segmentation,
including automatic thresholding, edge-based methods, and morphology-based
methods such as the watershed transform, often used to segment touching
objects.
Morphological operators enable you to detect edges, enhance contrast,
remove noise, segment an image into regions, thin regions or perform
skeletonization on regions. Morphological functions in Image Processing
Toolbox include:

Erosion and dilation.

Opening and closing.

Labelling of connected components.

Watershed segmentation.

Reconstruction.

Distance transform.

4.3.1.3 Image enhancement

Image enhancement techniques in Image Processing Toolbox enable us
to increase the signal-to-noise ratio and accentuate image features by modifying
the colours or intensities of an image. The following processes can be done:

Perform decorrelation stretching.

Perform histogram equalization.

Remap the dynamic range.

Perform linear, median or adaptive filtering.

14

Figure.4.1. Performing connected components analysis on an image

with non-uniform background intensity using MATLAB and image processing
toolbox.

The toolbox includes specialized filtering routines and a generalized

multidimensional filtering function that handles integer image types, offers
multiple boundary-padding options and performs convolution and correlation.
Predefined filters and functions for designing and implementing our own linear
filters are also provided.
4.3.1.4 Image Deblurring
Image deblurring algorithms in Image Processing Toolbox include
blind, Lucy-Richardson, Wiener and regularized filter deconvolution, as well as
conversions between point spread and optical transfer functions. These
functions help correct blurring caused by out-of-focus optics, movement by the
15

camera or the subject during image capture, atmospheric conditions, short

exposure time and other factors. All deblurring functions work with
multidimensional images.

Figure.4.2. Image of the sun using deblurring algorithms.

16

CHAPTER 5

PROJECT DESCRIPTION
5.1 PROBLEM DEFINITION
In the process of image acquisition and transmission, impulse noises
often cause serious degradation of the image quality. Among the various
filtering algorithms that have been proposed, the family of median filters is the
most popular and holds a dominant position in this area for its simplicity. The
most representative paradigm in this family is known as Switching Median
Filtering (SMF), which partitions the whole filtering process into two
sequential steps - noise detection and filtering. By utilizing the priority
knowledge obtained from the noise detection step, the filtering step could be
more targeted and does not need to touch those uncorrupted pixels. Obviously
the accuracy of the noise detection is critical to the final result. In terms of the
conflict of noise suppression and edge-preserving, our proposed method takes
directional difference method to handle it.
5.2 OVERVIEW OF THE PROJECT
As in image enhancement, the future ultimate goal of restoration
techniques is to improve an image in some predefined sense.
Restoration attempts to reconstruct or recover an image that has been
degraded by using a priori knowledge of the degradation phenomenon. Thus
restoration techniques are oriented toward modelling the degradation and
applying the inverse process in order to recover the original image.
In the proposed method corrupted pixels are detected by using the
minimum and maximum gray level intensity values in the sliding window and
the detected noisy pixels replaced by the median values of noise free pixels and
the directional difference filter is used for edge preserving and effective noise
suppression.
17

The intensity of the noisy pixel will be distinct from its nearest
surrounding pixels. Based on this criterion the proposed method focuses on
noisy pixel detection.
In the first stage, select the window size as 3x3, then the maximum and
minimum gray level values in the window are to identify the noisy pixels. In
next stage, these detected noisy pixels will be replaced by the median value of
the noise free pixels.
In last stage the corrupted pixels which are left unchanged in the
previous stage are restored by using directional difference filters, it also
provides better noise suppression and edge preservation. The proposed method
works well for highly corrupted images with noise densities as high as 60 %
with better PSNR value and improved visual quality.
5.3 DESCRIPTION OF THE ALGORITHM
The description of the algorithm has two stages, they are

Noise Detection Algorithm Along with Adaptive Median Filtering

Directional Difference Filtering Algorithm

5.3.1Noise Detection Algorithm Along With Adaptive Median Filtering

Step 1:
Initialize the window size w =3.
Step 2:
Compute the maximum, minimum and median value in window, they
can be explained as

respectively.

18

Step 3:
If

, then go to step 5. Otherwise, set N=N+1.

Step 4:
If
window size is

go to step 2. Otherwise, we replace

by

when the

Step 5:
If

, then

is noise free, else we replace

by

5.3.2 Directional Difference Filtering Algorithm

Step 1:
In this step, the absolute difference of the central pixel from the four
directional pixels in a 3 x 3 window is taken and stored in four sets. The set
giving the directional differences is given as

The four directions are

Step 2:
Where (0, 0) is the central pixel position. Then the sum of the four
directional differences are found as

19

Step 3:
Minimum among the sum are found and the median of the minimum
directional difference is found. Only noise free pixels and pre-processed pixels
in the concerned direction are considered for taking median.
)

k=1, 2, 3, 4

Thus the noisy pixels are replaced by the median value of the minimum
of the sum of four directional differences.

20

5.4 DATA FLOW DIAGRAM

Figure.5.1. Fundamental steps in digital image processing

21

5.5 DATABASE DESIGN

Standard testing images used in digital image processing are as follows

Figure.5.2. Standard testing images used in existing algorithms

(a) Lena, (b) Baboon, (c) Bridge, (d) Peppers.

22

(a)

(b)

(b)

(d)

Figure.5.3. Standard testing images used in proposed algorithm.

(a) Cameraman, (b) M83, (c) Eight, (d) Saturn.

23

CHAPTER 6

RESULTS AND DISCUSSION

6.1 SIMULATION RESULTS
The threshold values are optimized to get better restoration. The
proposed method restores highly corrupted images with noise densities as high
as 60% with improved PSNR values. The tuning parameters are optimized to
yield good result for all noise densities. Hence the fixed values of the tuning
parameters give consistent PSNR values and also MSE value is much reduced.
The proposed method is compared with the existing methods. Three
different test images are used and the corresponding PSNR values and the MSE
values are compared. It can be seen that in the proposed method the PSNR
values are consistent and gives good PSNR for highly corrupted noise with
better image quality. The noise mask is used for processing only the noisy pixels
whereas retaining the noise free pixels. Pixel by pixel processing enables better
image restoration and visual quality.
The Mean Square Error Value (MSE) of the proposed method is
comparatively less to the previous methods. It implies that the false and missed
detections are less increasing the percentage of right detection. The rate of
increase in wrong detection is less than the existing methods and this helps in
better restoration of images that are highly corrupted. Thus the results for
different methods and the proposed method for 60% corrupted Lena image of
512x512 images is shown below.

24

6.2 PERFORMANCE ANALYSIS

6.2.1 Comparison Table for 512x512 Lena Image
Noise
Density

DWMF

EDRIN

RDRIN

Proposed
Method

30

18.1087

25.5235

26.5900

30.1799

40

16.1872

22.4067

23.3969

29.5786

50

14.7031

19.8100

20.5384

28.7990

60

13.3538

17.4319

18.0601

27.6059

70

12.3060

15.4260

15.7887

25.9309

in %

Table.6.1 Comparison of PSNR values of proposed technique with

existing techniques for 512x512 Lena image
Noise

DWMF

EDRIN

RDRIN

Density in

Proposed
Method

%
30

1005.1

182.2764

142.5855

62.3867

40

1564.4

373.6041

297.4327

71.6505

50

2201.8

679.3297

574.4358

85.7386

60

3004

1174.6

1016.4

112.8477

70

3823.7

1864.1

1714.8

165.9563

Table.6.2. Comparison of MSE values of proposed technique with

existing techniques for 512x512 Lena image

25

Figure.6.1. Simulation results for 60% corrupted Lena image (a)

original image (b) noisy image (c) DWMF (d)EDRIN (e) RDRIN (f) Proposed
method

26

6.2.2 Comparison Table for 384 x 512 Peppers Image

Noise
Density

DWMF

EDRIN

RDRIN

Proposed
Method

30

18.1818

25.5927

27.0006

33.9014

40

15.9346

22.2179

23.1923

33.0055

50

14.1109

19.0566

20.0746

31.6128

60

12.6763

16.5281

17.2199

30.2138

70

11.4586

14.3798

14.7532

27.2174

in %

Table 6.3. Comparison of PSNR values of proposed technique with

existing techniques for 384x512 peppers image
Noise

DWMF

EDRIN

RDRIN

Density in

Proposed
Method

%
30

988.3335

179.3965

129.7239

26.4814

40

1658.2

390.2002

311.7806

32.5482

50

2523.4

808.0132

639.1722

45.8543

60

3511.2

1446.3

1233.4

61.9017

70

4647.5

2371.9

2176.5

123.4067

Table 6.4. Comparison of PSNR values of proposed technique with

existing techniques for 384x512 peppers image

27

Figure.6.3. Simulation results for 60% corrupted cameraman image

(a) original image (b) noisy image (c) DWMF (d)EDRIN (e) RDRIN (f)
Proposed method

29

Figure.6.2. Simulation results for 60% corrupted peppers image (a)

original image (b) noisy image (c) DWMF (d)EDRIN (e) RDRIN (f) Proposed
method

28

Figure.6.4. Simulation results for 60% corrupted reghuman image

(a) original image (b) noisy image (c) DWMF (d)EDRIN (e) RDRIN (f)
Proposed method

30

CHAPTER 7

CONCLUSION
From this proposed method we obtain novel edge preserving method
for salt and pepper denoising. The method is actually a combination of
switching median filter and directional median method. The advantages of this
proposed method are the directional initialization of filtering window size and
the precision of median value. Thus, no training or tuning is required. It is
ultimate filter for denoising salt and pepper noise. Simulation results show that
our proposed method performs better than the median filter and other
conventional edge preserving method, even at a high noise level. The PSNR is
high; MAE and Processing time is low. This proposed method is a fast method
in the algorithm of removing salt and pepper noise.
7.1 FUTURE ENHANCEMENTS
In the future, we will consider using wavelet thresholding for
the image denoising and also preserve the edges of the image. This
project can also extend for the removal of random valued impulse
noise.

31

CHAPTER 8

REFERENCES
[1]. W. Luo, Efficient removal of impulse noise from digital images,
IEEETrans. Consumer Electron., vol. 52, no. 2, pp. 523-527, 2006.
[2]. R. H. Chan, C. W. Ho, and M. Nikolova, Salt and pepper noise removal
by median-type noise detectors and detail preserving regularization, IEEE
Trans. Image Processing, vol. 14, no. 10, pp. 1479-1485, 2005.
[3]. A. Bovik, Handbook of Image and Video Processing, New York: Academic,
2000.
[4]. T. S. Huang, G. J. Yang, and G. Y. Tang, Fast two-dimensional median
filtering algorithm, IEEE Trans. Acoustics, Speech, Signal Process, vol. ASSP1, no. 1, pp. 1318, 1979.
[5]. T. Sun and Y. Neuvo, Detail preserving median based filters in image
processing, Pattern Recognition Letters, vol. 15, pp. 341-347, 1994.
[6]. F. Russo and G. Ramponi, A fuzzy filter for images corrupted by impulse
noise, IEEE Signal Process. Lett, vol. 3, no. 6, pp. 168-170, 1996.
[7]. S. Zhang and M. A. Karim, A new impulse detector for switching median
filter, IEEE Signal Process. Lett, vol. 9, no. 11, pp. 360-363,2002.
[8]. H. Ibrahim and N. S. P. Kong, Simple adaptive median filter for the
removal of impulse noise from highly corrupted images, IEEE Trans Consumer
Electronics, vol. 54, no. 4, 2008.
[9]. K. K. V. Toh and H. Ibrahim, Salt and pepper noise detection and
reduction using fuzzy swithcing median filter, IEEE Trans Consumer
Electronics, vol. 54. no. 4, 2008.
[10]. L. Bar and A. Brook, Deblurring of color images corrupted by impulsive
noise, IEEE Trans Image Processing, vol. 16, no. 4, 2007
32