International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 20 (2017) pp. 10169-10175
© Research India Publications. http://www.ripublication.com
10169
A Review Application Of Image Segmentation On Statistical Texture
Analysis
Sabiq Adzhani Hammam
1
, Tito Waluyo Purboyo
2
and Randy Erfa Saputra
3
1
College Student, Faculty of Electrical Engineering, Telkom University, Bandung, Indonesia.
2
Lecturer, Faculty of Electrical Engineering, Telkom University, Bandung, Indonesia.
3
Lecturer, Faculty of Electrical Engineering, Telkom University, Bandung, Indonesia.
Abstract
Image processing is becoming very important. A lot of data
from an image is then removed, which will be used for
identification or coverage. One of image processing technique
that is often used today is image segmentation. Image
segmentation is used as the first step to separate objects from
the background, so the object can be processed for other
purposes. Texture analysis is the development of science from
image segmentation. The first part of this paper describes the
segmentation of images in general, then continued with an
explanation of image segmentation in general. Then explain
the image segmentation. There are many application could be
used for image segmentation based on obtained method:
Segmentation On Cancer, Segmentation On Medical
Resonance Image, and Segmentation On Texture Batik. The
last part of this paper will explain the future work on image
segmentation.
Keywords: Image Segmentation, Texture Analysis, Medical
Resonance Image, Cancer, Batik
INTRODUCTION
Image segmentation is a stage that was first performed before
the image analysis stage in the process of image recognition
of a particular input. The function of image segmentation that
divides the image of its territories (region) based on the
similarity in its kind in the form of textures, colors, shapes,
and etc. Implementation of common image segmentation
among others can be found in facial recognition application,
fruit quality detection, canned food industry, batik industry,
and so forth. Many methods are used for image segmentation
in texture statistical analysis. From several methods, the
application of image segmentation could be applied in many
fields like medical image, cancer, and texture. Text -based
on images segmentation, grouping it using repeating patterns
of pixels in the image. The texture is a function of the spatial
variation in pixel intensity in the image. Based on its structure
the texture can be divided into macrostructural textures and
microstructures. The macrostructural texture has periodic
pattern repetition of an area, usually present in man-made
patterns and tends to be easily represented mathematically.
While the texture of microstructure has a loop that is not clear
so it is not easy to provide a comprehensive texture definition.
The difficulty in the process of grouping the usual image
textures is to define the texture boundaries since the
boundaries between textures are often unclear. So the process
of segmentation is needed to perform the separation or
grouping objects properly so the next result obtained good
results. In addition, there are two things that refer to image
segmentation based on statistical analysis of textures, namely:
(1) Analysis of surface roughness level and (2) pattern
structures analysis and orientation. These two things have
become the basis for the development of image segmentation
based on statistical analysis of textures. The purpose of image
segmentation is to recognize patterns of the image. Image
segmentation methods need development that is in it to be
able to process further using image segmentation based on
statistical analysis of texture. By using the development of the
results of image segmentation can be implemented in the
process of classification and identification of an image.
Concept of Texture Image Segmentation
In general, the concept of image segmentation is to partition
an image into regions or components that do not overlap.
These regions should be homogeneous and uniform like Gray
Levels and its texture. Segmentation is also commonly done
as a first step to implementing object classification. After
image segmentation is implemented, the features contained in
the image can be taken and classified.
The process of segmentation includes two activities: Pre-
processing and feature extraction based on statistical methods
contained in the texture. This process is a preliminary process
to get the feature value contained in the image taken and used
for the process of pattern recognition and classification.