See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312211120

A Study of Frei-Chen Approach for Edge Detection

Article · January 2017

CITATIONS READS

9 1,612

3 authors, including:

Robbi Rahim
Sekolah Tinggi Ilmu Manajemen Sukma, Medan, Indonesia
302 PUBLICATIONS   3,122 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

an Improvement and study about Decision Support System View project

Housing Urban Poor and Urban Revitalisation. KLN Grant Ministry of Research, Technology and Higher Education Republic of Indonesia View project

All content following this page was uploaded by Robbi Rahim on 11 January 2017.

The user has requested enhancement of the downloaded file.

 © 2017 IJSRSET | Volume 3 | Issue 1 | Print ISSN: 2395-1990 | Online ISSN : 2394-4099
Themed Section: Engineering and Technology

A Study of Frei-Chen Approach for Edge Detection

Dicky Apdilah1, Muhammad Yasin Simargolang2, Robbi Rahim*3
1
AMIK Intel Com Global Indo Kisaran, Jl. Graha Abdi Satya No 47-48, Kisaran, Sumatera Utara, Indonesia
2
Universitas Asahan, Jl. Jend. Ahmad Yani, Kisaran, Sumatera Utara, Indonesia
3
* Departement of Computer Engineering, Medan Institute of Technology, Jl. Gedung Arca No.52, Medan,
Indonesia

ABSTRACT
Digital image processing is a computational process that is widely used today starting from editing photos or also
the manipulation of the picture, one form of image processing is edge detection, edge detection in images is one
technique that can be used to mark parts into detail of the picture, either a blurred image due to error or the effect of
the image acquisition process, in this study using the Frei-Chen algorithm to perform edge detection image in order
to know the borders of the picture.
Keywords : Digital Image Processing, Image Processing, Frei-Chen, Edge, Edge Detection

I. INTRODUCTION use for edge detection, for example, is a method Sobel,

Canny, Prewitt, Frei-Chen and Laplacian [4] [8] [9]
Digital image is a picture of an object that is an analog [10]. This study uses Frei-Chen, Frei-Chen an edge
form of video signals, or digital can also directly stored detection method using the Frei-Chen mask which
on a storage media such as flash and various other contains the base vector calculation to be applied to the
storage media [1] [2]. The use of digital images has image [11]. Frei-Chen pointed a simple edge detector
increased because of the advantages possessed by the best does edge detection, followed by thinning and
digital image, among other things ease in obtaining linking processes to optimize the margins [1] [8]. This
imagery, image processing, and others. But not all research tries to apply the method Frei-Chen to perform
digital images have a pleasing visual appearance of the edge detection in digital images with the help of tool
human eye [1] [2]. Matlab.

One form of image processing that can use is the edge II. METHODS AND MATERIAL
detection [1] [3]. Edge detection is common in digital
image processing because the first step in image The image is the visual representation of an object [1]
segmentation, image segmentation is used to present the [2], the output of a system such as optical data recording
objects contained in a picture [3] [4] [5]. Edge detection in the form of photos or also the analog video signal
function to identify the boundaries of an object on the form as the image on the monitor and also directly
image overlap [6]. Therefore, when the outline of the stored on a storage medium [2]. Images grouped into
picture can identify accurately, all objects can be found, two parts, namely the still image and a moving image.
and basic properties such as area, shape, and size of the The still image is a single image that is not moving.
object can be measured [6] [7]. The edges of the image Instead, the moving image is a series of a still image is
are the position where the intensity of the pixels of the displayed in a sequence, thus giving the impression to
picture changed from low value to high value or vice the eye as a moving picture [2].
versa [7]. Currently, there are several methods that can

IJSRSET17318 | Received : 03 Jan 2017 | Accepted : 11 Jan 2017 | January-February-2017 [(3)1: 59-62 ]
59
 Edge Detection

Edge detection is the first step to cover the information

in the image, the edges characterize the boundaries of
the object, therefore, useful to the process edge
segmentation and identification of objects in the picture.
Interest edge detection operation is to improve the
appearance of the boundary line of an area or object in
the image [4] [5] [6] [9] [11]. Figure 1. Sample Image

Convolution from the picture above are assumed grayscale values

obtained from the use Matlab tool is as follows:
Implementation of a filter on an image used a technique
called convolution [2], convolution is express in the
form of a matrix n, where each element is called
coefficient convolution matrix [1] [9] [10]. Process
convolution kernel works by shifting pixels per pixel,
( )
which results stored in the new matrix. Here is an
example of the convolution that occurs between image f The next is to determine the kernel that is used to detect
(x, y) 5x5 with a 3x3 kernel shown in the following the image, this kernel function to be used as a base
figure. count by using algorithms Frei-Chen here is the kernel

( )
[ √ √ ]
( )
After determining the value of the next kernel image
( ) ( ) and calculate the value of the convolution of the picture.

TABLE I. First Values

Mark. (Dot) indicates the position (0,0) of the
convolution kernel to get the results that occur between 55 63 77 80 129
image and kernel. 22 94 87 27 26
107 154 76 30 27 x [ √ √ ]
Frei-Chen Algorithm 108 252 220 75 30
20 111 185 51 10
Edge detection using Chen Frei implemented by
mapping a vector mask intensity using linear after knowing the value of the kernel image and the next
transformations, and then detect the intensity of the is to do the calculation as follows:
edge based on the angle between the vectors are
projected into space the edges. Frei-Chen Mask is a (-1 x 55) + (0 x 63) + (1 x 77) + ( √2 x 22) + (0 x 94) +
unique mask, which all contain a base vector mask if (√2 x 87 ) +(-1 x 107) +(0 x 154) +(1 x 76) = 82
implemented on a 3X3 image area represented by nine
Frei-Chen mask [2] [5] [8] [9]. From the calculation a new value is made from an
existing matrix, the results are as follows:
III. RESULTS AND DISCUSSION

The process of testing Frei-Chen algorithm in edge

detection performed in the following figure

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

60
 TABLE II. New Value (-1 x 77) + (0 x 80) + (1 x 129) + ( √2 x 87) + (0 x 27)
+ (√2 x 26 ) +(-1 x 76) + (0 x 30) +(1 x 27) = -82.4

82 TABLE VI. New Value after Shift Value Process 3

82 -200.8 82.4

The next process is the kernel shift one pixel to the

right, then calculate the pixel value at the position (0,0)
of the kernel
Next, slide the kernel one pixel to the bottom and start
TABLE III. Shift Value Process 2 again doing the convolution of the left side of the
image. Every time convolution, sliding kernel one pixel
55 63 77 80 129 to the right
22 94 87 27 26
107 154 76 30 27 x [ √ √ ] TABLE VII. Next Pixel Value Process 4
108 252 220 75 30
20 111 185 51 10 55 63 77 80 129
22 94 87 27 26
after knowing the value of the kernel image and the next 107 154 76 30 27 x [ √ √ ]
is to do the calculation as follows: 108 252 220 75 30
20 111 185 51 10
(-1 x 63) + (0 x 77) + (1 x 80) + ( √2 x 94) + (0 x 87) +
(√2 x 27 ) +(-1 x 154) + (0 x 76) +(1 x 30) = -200.8 after knowing the value of the kernel image and the next
is to do the calculation as follows:
TABLE IV
New Value after Shift Value (-1 x 22) + (0 x 94) + (1 x 87) + ( √2 x 107 + (0 x 154)
+ (√2 x 76 ) +(-1 x 108) + (0 x 252) +(1 x 220) = 133.6
82 -200.8
TABLE VIII. New Value after Shift Value Process 4

82 -200.8 82.4
The next process is the kernel shift one pixel to the 133.6
right, then calculate the pixel value at the position (0,0)
of the kernel

TABLE V. Shift Value Process 3 do the shift and pixel displacement up to the whole
kernel convolution matrix values calculate, so that the
55 63 77 80 129 outcome of the process were calculated as follows
22 94 87 27 26
107 154 76 30 27 x [ √ √ ] TABLE IX. Final Result
108 252 220 75 30
20 111 185 51 10 55 63 77 80 129
22 82 -200.8 82.4 26
after knowing the value of the kernel image and the next 107 133.6 -417.6 -319.6 27
is to do the calculation as follows: 108 290.8 -413.8 -490 30
20 111 185 51 10

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

61
 above values a the final value calculation process by Communication Engineering, vol. 2, no. 1, pp.
using algorithms Frei-Chen, from the calculation, the 994-999, 2013.
result of edge detection of Figure 1 is as follows: [7]. P. Zhou, W. YE, Y. XIA and Q. WANG, "An
Improved Canny Algorithm for Edge Detection,"
Journal of Computational Information Systems,
vol. 7, no. 5, pp. 1516-1523, 2015.
[8]. R.-H. PARK, "A Fourier interpretation of the
Frei-Chen edge masks," Elsevier, vol. 11, no. 9,
pp. 631-636, 1990.
[9]. Muthukrishnan and M.Radha, "EDGE
DETECTION TECHNIQUES FOR IMAGE
SEGMENTATION," International Journal of
Figure 2. Result Image Computer Science & Information Technology,
vol. 3, no. 6, pp. 259-267, 2011.
IV. CONCLUSION [10]. M. A. Zahhad, R. R. Gharieb, S. M.Ahmed, and
A. A. El-Baset Donkol, "Edge Detection with a
Preprocessing Approach," Journal of Signal and
Frei-Chen algorithm testing at the image edge detection
Information Processing, vol. 5, pp. 123-134,
can be done well with good results, from the count of
2014.
kernel algorithms Frei-Chen conducted the process is
[11]. Y.Ramadevi, T.Sridevi, B.Poornima and
also easy to do, and the application is made using a tool
B.Kalyani, "SEGMENTATION AND OBJECT
like Matlab image processing is relatively easy.
RECOGNITION USING EDGE DETECTION
TECHNIQUES," International Journal of
Computer Science & Information Technology,
V. REFERENCES
vol. 2, no. 6, pp. 153-161, 2010.
[12]. M. Al-khassaweneh, "An Image Encryption
[1]. T. Acharya and A. K.Ray, Image Processing Model Based on Using Frei-Chen Bases," in
Principles, and Applications, Canada: John Wiley International Conference on Modeling,
& Sons, 2005.
Simulation and Applied Optimization, 2011.
[2]. J. C. Russ, The Image Processing Handbook The
Sixth Edition, Florida: CRC Press, 2011.
[3]. Nisha, R. Mehra and L. Sharma, "Comparative
Analysis of Canny and Prewitt Edge Detection
Techniques used in Image Processing,"
International Journal of Engineering Trends and
Technology (IJETT), vol. 28, no. 1, pp. 48-53,
2015.
[4]. S. S. Al-Amri, N.V. Kalyankar and Khamitkar
S.D, "IMAGE SEGMENTATION BY USING
EDGE DETECTION," International Journal of
Computer Science and Engineering (IJCSE), vol.
2, no. 3, pp. 804-807, 2010.
[5]. R. Maini and H. Aggarwal, "Study and
Comparison of Various Image Edge Detection
Techniques," International Journal of Image
Processing, vol. 3, no. 1, pp. 1-12, 2009.
[6]. S. Saluja, A. K. Singh and S. Agrawal, "A Study
of Edge-Detection Methods," International
Journal of Advanced Research in Computer and

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

62

View publication stats