INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.2 Issue.9, Pg.: 107-112
September 2014
INTERNATIONAL JOURNAL OF
RESEARCH IN COMPUTER
APPLICATIONS AND ROBOTICS
ISSN 2320-7345
FUZZY LOGIC TECHNIQUE FOR NOISE
REMOVAL IN EDGE DETECTION METHODS
1
1
Dr.T.Karthikeyan, 2N.P.Revathy
Associate Professor, P.S.G. College of Arts and Science, Coimbatore, Email: t.karthikeyan.gasc@gmail.com
2
Research scholar, Karpagam University, Coimbatore, Email: np.revathy@yahoo.in
Abstract
Image Edge detection significantly reduces the amount of data and filters out useless information, while preserving
the important structural properties in an image. This paper presents a fuzzy rule base algorithm which is capable of
detecting edges efficiently from the gray scale images. The following paper introduces such operators on hand of
computer vision application. In the proposed algorithm, edginess at each pixel of a digital image is calculated using
three 3 3 linear spatial filters i.e. low-pass, high-pass and edge enhancement (Sobel) filters through spatial
convolution process. Finally We Have compared results of the proposed algorithm with other algorithms such as
Sobel, Robert, and Prewitt. Experimental results show the ability and high performance of proposed algorithm.
Keywords: Edge Detection, WSVM, Fuzzy cognitive map, FNN Noise Removal, Digital Image Processing.
1. INTRODUCTION
Edge detection represents an extremely important step facilitating higher-level image analysis and therefore remains
an area of active research, with new approaches continually being developed
Edge detection is an indispensable step in the computer vision and object recognition, because the most fundamental
characteristic for image recognition is the edges of an image. Edges are boundaries between different textures. Edge
also can be defined as discontinuities in image intensity from one pixel to another [1]. The edges for an image are
always the important characteristics that offer an indication for a higher frequency. The goal of edge detection is to
convert a 2D image into a set of curves. The salient features are expected to be the boundaries of objects that tend to
produce sudden changes in the image intensity. They can show where shadows fall in an image or any other distinct
change in the intensity of an image. The quality of edge detection is highly dependent on lighting conditions, the
presence of objects of similar intensities, density of edges in the scene.
Since different edge detectors work better under different conditions, the objective of this paper is to identity the
suitable edge detector for the IR images with the noise present in it [2]. Moreover, the aim is to find the good edge
detector for the filtered images and the respective filter used for the identification of the same. The reason for
choosing IR image is that quality of the IR images differs from the normal images. The edge detection operators
such as the Sobel, Prewitt, Log and Canny are commonly used in the process of the edge detection. The noise taken
for this study is Salt and Pepper noise a type of Impulsive noise, where the noise value may be either the minimum
or maximum of the dynamic gray scale range of the image.
N.P.Revathy et al
Page 107
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.2 Issue.9, Pg.: 107-112
September 2014
Common traditional edge extraction algorithms such as Canny use a constant window that can be mixed with some
smoothing filters. They need high quality values for their parameters in order to reach extraction efficiency [3, 4].
Despite simplicity, low computational cost and the fact that these parameters are known in high degree of quality
based on experiences during last years, but they are still dependent to lightening conditions, noise etc. Lack of any
of these dependencies could result in fail of these methods. In addition, using a constant parameter all over the
image can result as discontinuity in edges ant this discontinuity in edges is one of the most important weaknesses in
such algorithms. Some methods try to extract special edges by applying transformations such as Hough transform
but all edges don’t meet required conditions. Due to lack of information, using hybrid techniques usually leads the
process to fail. .In this paper, fuzzy logic based approach to edge detection in digital images is proposed. Firstly, for
each pixel in the input image edginess ‘measure is calculated using three 3 x3 linear filters after which three fuzzy
sets characterized by three Gaussian membership functions associated to linguistic variable Low [11], Medium and
High were created to represent each of the edge strengths. The second phase involves application of fuzzy inference
rule to the three fuzzy sets to modify the Membership values in such a way that the fuzzy system out-put (edge) is
high only for those pixels belonging to edges in the input image. The Last step is final pixel classification as edge or
non-edge using Mamdani de-fuzzification method.
2. APPLICATION OF FUZZY LOGIC BASED EDGE DETECTION
Fuzzy logic represents a powerful approach to decision making [Zadeh 1965, Kaufmann 1975 and Bezdek1981].
Since the concept of fuzzy logic was formulated in 1965 by Zadeh, many researchers have been carried out on its
application in the various areas of digital image processing such as image quality assessment, edge detection, image
segmentation, etc[3]. Many techniques have been suggested by researchers in the past for fuzzy logic-based edge
detection [Cheung and Chan 1995, Kuoet al. 1997, ElKhamy et al. 2000]. In [Zhao,2001].proposed an edge
detection technique based on probability partition of the image into 3-fuzzy partitions (regions) and the principle of
maximum entropy for finding the parameters value that result in the best compact edge representation of images. For
example in [6], adjacent points are assumed as 3×3 sets around the concerned point. By predefining Member-ship
function to detect edges. In these rules discontinuity in the color of different 3×3 sets, edges are extracted. It uses 5
fuzzy rules and predefined membership function to detect edges. In these rules discontinuity of adjacent point
around the concerned point are investigated. If this difference is similar to one of predefined sets, the pixel is
assumed as edge [6].
Image
Image fuzzification
Fuzzy inference system
Image de-fuzzificational
Gradient
Entropy
Variance
Fuzzy neural network based Fuzzy
cognitive map for edge detection
Edge
Non-edge
Figure 1 : Flow diagram of proposed edge detection
N.P.Revathy et al
Page 108
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.2 Issue.9, Pg.: 107-112
September 2014
3. PROPOSED METHODOLOGY
In this paper, at first an input image is pre-process to accentuate or remove a band of spatial frequencies and to
locate in an image where there is a sudden variation in the grey level of pixels. For each pixel in the image edge
strength value is calculated with three 3×3 linear spatial filters i.e. low-pass, high-pass and edge enhancement filters
(Sobel) through spatial convolution process. In carrying out a 3×3 kernel convolution, nine convolution coefficients
called the convolution mask are defined and labelled as seen below.
P1
P2
P3
P4
P5
P6
P7
P8
P9
Fig. 2. Applied mask to compute standard deviation
I ( x, y )= g( k, l ) *I ( x, y )
where: g (k, l) = convolutional kernel
I (x,y) = original image
I’(x,y) = filtered image
2N + 1 = size of convolutional kernel
If SD value of a pixel being equal to k1 , and gradient be equal to k2 , the fuzzy rules are defined as the following :
1-If k1 in SDL & k2 in GL then P Edge classified to EL
2-If k1 in SDL & k2 in GM then P Edge classified to EL
3-If k1 in SDL & k2 in GH then P Edge classified to EM
4-If k1 in SDM & k2 in GL then P Edge classified to EL
5-If k1 in SDM & k2 in GM then P Edge classified to EM
6-If k1 in SDM & k2 in GH then P Edge classified to EH
7-If k1 in SDH & k2 in GL then P Edge classified to EM
8-If k1 in SDH & k2 in GM then P Edge classified to EH
9-If k1 in SDH & k2 in GH then P Edge classified to EH
After the image data are transformed from gray-level plane to the membership plane (fuzzification), appropriate
fuzzy techniques modify the membership values. This can be a fuzzy clustering, a fuzzy rule-based approach, a
fuzzy integration approach and so on, [29]. In this paper a novel FIS method based on fuzzy logic reasoning strategy
is proposed for edge detection in digital images without determining the threshold value or need training algorithm.
This study is assaying all the pixels of the processed image by studying the situation of each neighbor of each pixel.
The condition of each pixel is decided by using the floating 3x3 mask which can be scanning the all grays. In this
location, some of the desired rules are explained.
After the image data are transformed from gray-level plane to the membership plane (fuzzification), appropriate
fuzzy techniques modify the membership values. This can be a fuzzy clustering, a fuzzy rule-based approach, a
fuzzy integration approach and so on, [29]. In this paper a novel FIS method based on fuzzy logic reasoning strategy
is proposed for edge detection in digital images without determining the threshold value or need training algorithm.
This study is assaying all the pixels of the processed image by studying the situation of each neighbor of each pixel.
The condition of each pixel is decided by using the floating 3x3 mask which can be scanning the all grays. In this
location, some of the desired rules are explained
N.P.Revathy et al
Page 109
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
(a)
(b)
(c)
Vol.2 Issue.9, Pg.: 107-112
September 2014
(d)
Fig. 3 (a) Original images, (b) Sobel operator results, (c) Kirsch operator results, (d) Proposed fuzzy edge detection
algorithm results.
The proposed fuzzy edge detection method was simulated using MATLAB on different images, its performance are
compared to that of the Sobel and Kirsch operators. Samples for a set of four test images are shown in Fig. 3(a). The
edge detection based on Sobel and Kirsch operators using the image processing toolbox in MATLAB with threshold
automatically estimated from image‘s binary value is illustrated in Fig. 3(b) and 3(c). The sample output of the
proposed fuzzy technique is shown in Fig. 3(d). The resulting images generated by the fuzzy method seem to be
much smoother with less noise and has an exhaustive set of fuzzy conditions which helps to provide an efficient
edge representation for images with a very high efficiency than the conventional gradient-based methods (Sobel and
Kirsch methods). Now I am applying this method on color image with the help of two degree angle 45 and 30
degree.
Fig. 4. Color Image with 45º
N.P.Revathy et al
Page 110
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.2 Issue.9, Pg.: 107-112
September 2014
4. EXPERIMENTAL RESULTS
The proposed edge detection method is simulated using MATLAB on different images. It is observed that this
proposed system provide much more distinct marked edges as compared to edge detection algorithms such SVM
and weighted SVM. The performance metrics used for analyzing the proposed metrics are MSE, PSNR.
Peak Signal to Noise Ratio
The ratio between the maximum possible powers to the power of corrupting noise is known as Peak Signal to Noise
Ratio. It affects the fidelity of its representation .It can be also said that it is the logarithmic function of peak value of
image and mean square error.
Mean Square error
Mean square error (MSE) of an estimator is to quantify the difference between an estimator and the true value of the
quantity being estimated.
Methodology
MSE
PSNR
SVM
0.3259
34.258
Number of
edges
detected
989
WSVM
0.2369
37.259
2045
HFCM-FNN
0.2145
40.256
3125
Figure 5: MSE comparison for edge detection methods
Figure 6 : Number of edges detected results for edge detection methods
N.P.Revathy et al
Page 111
INTERNATIONAL JOURNAL OF RESEARCH IN COMPUTER APPLICATIONS AND ROBOTICS
www.ijrcar.com
Vol.2 Issue.9, Pg.: 107-112
September 2014
5. CONCLUSION
In this paper, effective fuzzy logic based edge detection has been presented. This technique uses the edge strength
information derived using three masks to avoid detection of spurious edges corresponding to noise, which is often
the case with conventional gradient-based techniques. The three edge strength values used as fuzzy system inputs
were fuzzified using Gaussian membership functions. Fuzzy if-then rules are applied to modify the membership to
one of low, medium, or high classes. This work presents improved edge detector that uses the Hybrid Fuzzy
Cognitive Map based Fuzzy Neural Network(HFCM-FNN) to perform the pixel classification between edge and no
edge.
REFERENCES
[1] Canny, J.F., A computational approach to edge Intelligence, 8(6), 1986, pp. 679-698.
[2] Laligant, F. Truchete, and J. Miteran , Ed gedetection by multiscale merging, in Proceedings of the IEEESP International Symposium on Time-Frequency and Time- Scale Analysis, pp. 237-240, IEEE,
LosAlamitos,CA,1994.
[3] M. Sun and R. J. Sclabassi, Sym metric wavelet edge detector of the minimum length, In International
Conference on Image Processing, pp. 177-180. IEEE, Los Alamitos, CA, 1995.
[4] W. Chen and S. Acton, "Morphological pyramids for multiscale edge detection,"IEEE Southwest
Symposium on Image Analysis and Interpretation, pp. 137-141, 5-7 April 1998.
[5] S. Lu, Z. Wang, and J. Shen, "Nero-fuzzy synergism to the intelligent system for edge detection and
enhancement, "the journal of the Pattern Recognition society, accepted 1October 2002.
[6] G. Mansoori and H. Eghbali, "Heuristic edge detection using fuzzy rule-based classifier,"Journal of
Intelligent and Fuzzy Systems, pp. 457 – 469, Volume 17, Number 5 /2006.
[7] H. Nezamabadi-pour, S. Saryazdi, and E. Rashedi, "Edge detection using ant algorithms,"Soft Computing ,
vol.10, no.7, pp. 623-628, 2006.
[8] H. Wong and L. Guan, "A Neural Learning Approach for Adaptive Image Restoration Using a Fuzzy
Model-based Network Architecture,"IEEE Trans, Neural Network 12(3) 516-531, 2001.
[9] C. Looney, "Nonlinear Rule-based Convolution for Refocusing,"RealTime Imaging , (6) 29-37,2000.
[10] H. M. Lozoya, D. M. Rodriguez, F. J. Romero, and H. Tawfik, "Handoff Algorithms Based on Fuzzy
Classifiers," IEEE Trans, Vehicular Technol,49 (6) 2286-2294,2000.
[11] K. Cheung and W. Chan, "Fuzzy One –Mean Algorithm forEdge Detection, "IEEE Inter.Conf. On Fuzzy
Systems, pp. 2039- 2044, 1995.
[12] Huaibin Wang , Li Wang ,Application of Improved Fuzzy Cognitive Map Based on Fuzzy Neural Network
in Intrusion Detection , Journal of Information & Computational Science, Vol.10,No. 1,pp.271–278, 2013.
N.P.Revathy et al
Page 112