A Comparative Study of Various Edge Enhancement Filters in Spatial Domain

 

A.V.S.N. Murty¹, B.N. Jagadesh², K. Bhagavan², S. Satyanarayana²

¹School of Advanced Sciences, Department of Mathematics, VIT University, Vellore-632014, TN, India.

²Department of CSE, K L University, Vaddeswararam, Guntur (Dt.), Andhra Pradesh, India.

 

ABSTRACT:

Edge detection plays an important role in many Image processing applications. In any Image Segmentation technique edges are primary features. Edge detection refers to the process of identifying and locating sharp discontinuities in an image. So, edge detection is an important part in image analysis and it is also playing a key role of solving many complex problems. In this paper, our main aim is to compare the various edge detection algorithms for image segmentations.

 

 

 

INTRODUCTION:

Image segmentation is one of the important parts of image analysis. Image segmentation is a process of dividing an image into group of homogeneous pixels. The major applications that uses image segmentation either at pre-processor level or at advanced level. Segmentation algorithms are area oriented instead of pixel oriented. Segmentation methods are categorized on the basis of two properties discontinuity and similarity. Several algorithms and techniques have been developed for image segmentation. Different traditional segmentation techniques such as threshold based techniques, histogram based techniques, edge based techniques and region based techniques are available in literature but still it remains as a challenge for computer vision to satisfy global needs and hence selection of proper techniques is very important as per application. Researchers have been extensively worked over this problem and proposed that Edge-based segmentation is one of the widely used techniques for image segmentation [1].

 

In edge-based approach, the partitions or sub-division of an image is based on some abrupt changes in the intensity level of images [2]. In edge detection noise also influencing, this degrades the image quality for segmentation of image. Several factors are responsible for noise in the image during image acquisition or transmission. Various types of noises are usually present in the image like Impulse noise, Amplifier noise, Quantization noise etc. In this paper, our main aim is to execute different edge detection techniques such as Sobel, Prewitt, Canny, Kalman and Gabor filters with noisy/noise less images and compare with each other. This paper is organized as follows. Section 2 is for the purpose of providing some information about edge detection for image segmentation. Section 3 is focused on different edge detection methods. Section 4 explains threshold techniques and Section 5 concentrates on comparison of various edge detection methods. Section 6 presents the conclusion.

 

Edge Detection for Image Segmentation:

In the edges, it can sufficiently modify the limited part of the images. Detection is the action or the process of identifying the presence of something to be cancelled. The edge it is used for the important features can be extracted from the edges of an image. According to the intensity values the edges can be modelled in the following Fig.1

 

Fig.1.Type of Edges (a) Step Edge (b) RampEdge (c) Line Edge (d) Roof Edge

 

An Edge in an image is a significant local change in the image intensity, usually associated with a discontinuity in either the image intensity or the first derivative of the image intensity. Discontinuities in the image intensity can be either Step edge, where the image intensity abruptly changes from one value on one side of the discontinuity to a different value on the opposite side, or Line Edges, where the image intensity abruptly changes value but returns to the starting value within some short distance [3]. However, Step and Line edges are rare in real images. Because of low frequency components or the smoothing introduced by most sensing devices, sharp discontinuities rarely exist in real signals. Step edges become Ramp Edges and Line Edges become Roof edges, where intensity changes are not instantaneous but occur over a finite distance [4]. Illustrations of these edge shapes are shown in Fig.1.

 

A. Steps in Edge Detection:

Edge detection contains three steps namely Filtering, Enhancement and Detection. The overview of the steps in edge detection is as follows.

 

Filtering:

Images are frequently corrupted by noise. Some common types of noise are salt and pepper noise, impulse noise and Gaussian noise. Salt and pepper noise contains random occurrences of both black and white intensity values. However, there is a trade-off between edge strength and noise reduction [5].

 

Enhancement:

In order to facilitate the detection of edges, it is essential to determine changes in intensity in the neighbourhood of a point. Enhancement emphasizes pixels where there is a significant change in local intensity values and is usually performed by computing the gradient magnitude [6].

 

Detection:

Many points in an image have a non-zero value for the gradient, and not all of these points are edges for a particular application. Therefore, some method should be used to determine which points are edge points. Frequently, thresholding provides the criterion used for detection [7]. Edge detection is the process that which it can be used for the identifying and locating the large gaps which had presented in the image. Edge detection is used to detect the edges by using the smoothing or sharpening filter. The discontinuities are unexpected changes in pixel intensity that take some area of the image that is composed to describe the features of the boundaries of objects. The most commonly used discontinuity based edge detection techniques are Laplacian-based; Roberts based, sobel-based, krish-based, robinson-based, marrr-hildreth based, and log edge detection. Among these techniques Sobel, Prewitt and Canny edge detectors are have some special characteristics.

 

EDGE DETECTION METHODS:

There are various techniques available for edge segmentation to perform detection of edges in the image. Here, we are discussing some of the techniques which shows better performance compared to other techniques. They are:

 

Sobel Based Edge Detection:

The Sobel edge detection method is introduced by Sobel in 1970. The Sobel method of edge detection for image segmentation finds edges using the Sobel approximation to the derivative. It precedes the edges at those points where the gradient is highest. The Sobel technique performs a 2-D spatial gradient quantity on an image and so highlights regions of high spatial frequency that correspond to edges. In general it is used to find the estimated absolute gradient magnitude at each point in “n” input gray scale image. The Sobel operator consists of a pair of 3×3 convolution kernels as shown in Table.1

 

These convolution kernels are separately applied tothe edges on images running vertically and horizontally, and produce separate measurements ineach orientation(Gx & Gy).These vertical and horizontal results are combined together to find the absolute magnitude of the gradient at each point and orientation of that gradient. Typically, an approximate magnitude is computed using:

 

|G| = |Gx| + |Gy|

 

As we know that, this filter is showing poor performances with resolution images and does not consider high frequency content of images.

 

Prewitt Edge Detector:

The Prewitt edge detector is an appropriate way to estimate the magnitude and orientation of edge. Although differential gradient edge detection need a rather time consuming calculation to estimate the orientation from magnitudes in the x &y directions.

 

The above diagram shows the masks used by a Prewitt edge detector digitally the first Gx and Gy. The parameters of this function are same as the Sobel parameters. The Prewitt detector is somewhat easier to execute computationally than the Sobel detector, but it generates slightly noise results. The syntax employed by prewitt edge detector is[g,t]=edge(f, “prewitt”, T, dir).

 

Canny Edge Detector:

The Canny edge detection operator was developed by John F. Canny in 1986 and uses a multi-stage algorithm to detect a wide range of edges in images. Canny also produced a computational theory of edge detection explaining why the technique works. The canny edge detector is regarded as one of the best edge detectors currently in use; it ensures the good noise immunity and at the same time detects the edge points with minimum error. In canny algorithm there are some steps as follows:

 

Smoothing:

blurring the image to remove noise by convolving the image with the Gaussian filter.

 

Finding Gradients:

The edges should be marked where the gradient of the image has large magnitude.

Non-Maximum Suppression:

only local maxima should be marked as edges

 

Double Thresholding:

potential edges are determined by thresholding. There are two threshold levels, th, high and tl, low where th>tl. Pixel values above the value are immediately classified as edges. Canny edge detection algorithm is computationally more expensive compared to sobel, prewitt, laplacian and Roberts operator.

 

Kalman Edge Detector:

A Kalman filter is an optimal solution for edge detection techniques [9]. The Kalman filter addresses the general problem of trying to estimate the state of a discrete-time controlled process that is governed by the linear stochastic difference equation [10]

The random variables and represent the process and measurement noise respectively. The kalman filter gives good results due to optimality and structure and it is too convenient for dynamic images.

 

THRESHOLDING:

Thresholding technique computes very fast and inexpensive, it is the oldest segmentation method and is still used in normal applications. This is the simplest way of segmentation methods based on image regions [8]. Principle of thresholding techniques is based on the characteristics of the image. By using this technique, regions are classified based on the range values, which is applied to the intensity values of image pixels. In brightness threshold, all the pixels brighter than a specified dark even are taken as 1 and unwanted background as 0. Basically thresholding techniques classified into two types. They are local thresholding-here; we select different values for different regions. Global thresholding-In this method, we select only one threshold value for whole image. For segmenting complex images multilevel thresholding is required. Various types of thresholding methods are available they are 1.ostu thresholding, 2.histogram thresholding and 3. Iterative thresholding Here we consider the histogram thresholding for segmenting the images. Histogram thresholding is very efficient when compared to other threshold segmentation methods because they typically require only one pass through the pixels. In this technique, a histogram is computed from all of the pixels in the image, and the peaks and valleys in the histogram are used to locate the  clusters in the image. A refinement of this technique is to  recursively apply the histogram-seeking method to clusters in the image in order to divide them into smaller clusters. This is repeated with smaller and smaller clusters until no more clusters are formed. In this paper, we have done a comparison of various edge detection techniques on different types of noisy images. Finally, kalman filter performs better than that of existing algorithms.

 

Experimental Results and Analysis:

In this section, the efficiency of the various edge detection techniques is evaluated with some standard images. The following figure shows the original and segmented images with different edge detection techniques such as Sobel, Prewitt, and Canny and Kalman operators.

 

The above figure also shows that Kalman filter is better than that of remaining edge detection techniques and performance of the edge detection techniques are measured by Mean Square Error (MSE). MSE indicate s the average difference of the pixels throughout the image.

 

Table.1 Mean Square Error for Edge detection techniques

 

The above table also reveals that Kalman edge detection is having less mean square error. So that kalman filter performs better than that of existing methods.

CONCLUSION:

In this paper, we mainly focused on the study of the segmentation performance for the filters approach on the edge detection for image segmentation. We conclude that Kalman edge detector performs better and producing some better results compared to the other techniques.

 

REFERENCES:

1.     N. Senthilkumaran, R. Rajesh, "Edge Detection Techniques for Image Segmentation and A Survey of Soft Computing Approaches", International Journal of Recent Trends in Engineering, Vol.1, No.2, PP.250-254, May2009.

2.     Lakshmi,S, V.Sankaranarayanan A Study of edge detection techniques for segmentation computing approaches”, Computer Aided Soft Computing Techniques for Imaging and Biomedical Applications,35-41, 2010.

3.     M.Abdulghafour,”ImagesegmentationusingFuzzylogicandgeneticalgorithms”, Journal of WSCG, vol.11, no.1, 2003.

4.     Metin Kaya,”Image Clustering and Compression Using an Annealed Fuzzy Hopfield Neural Network”, International Journal of Signal Processing ,pp.80-88, 2005.

5.     N. Senthilkumaran, R. Rajesh, “A Study on Split and Merge for Region based Image Segmentation”, Proceedings of UGC Sponsored National Conference Network Security(NCNS-08) ,pp.57-61, 2008.

6.     Xian Bin Wen, Hua ZhangandZe Tao Jiang, “Multiscale Unsupervised Segmentation of SAR Imagery Using the Genetic Algorithm”, Sensors, vol.8,pp.1704-1711, 2008.

7.     Mantas Paulinas, AndriusUsinskas, “A Survey of Genetic Algorithms Applicatons for Image Enhancement and Segmentation”, Information Technology and Control, Vol.36, No.3, pp.278-284, 2007.

8.     Orlando J. Tobias, RuiSeara, ”Image Segmentation by Histogram Thresholding Using Fuzzy Sets”, IEEE Transactions on Image Processing,Vol.11,No.12,2002,pp.1457-1465.

9.     Mahbub Murshed, Md. Hasanul Kabir, Oksam Chae, “ Moving object Tracking- An edge segment based approach”, International  Journal of Innovative Computing, Information and Control, Vol.7, Numer.7(A), 2011.

10.   Vidyashree. K, K. V. Mahendra Prasanth, “Design of a vehicular tracking system using combined CED algorithm and Fuzzy logic”, International Journal of Research in Engineering and Technology, Vol.5, Issue.6, 2016.

 

 

Accepted on 08.11.2016        © RJPT All right reserved

Research J. Pharm. and Tech 2016; 9(12):2403-2406.