Sistem pemprosesan imej untuk pengesananbarah payudara
Loading...
Date
2002
Authors
Sulaiman, Siti Noraini
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Breast cancer is well known as a mortal disease. In United States, it is reported as the
second leading cause of death in women after lung cancer. However, if the cancer can
>c detected <1;t its earliest s~age, the .Qumber. of de?-th can be reduced. Mammogram and
dtrasound are the best two methods usually used for breast cancer detection.
Jnfortunately, problems normally arise where the physician could not make a good
.lecision just because of the images are blurred, not cleared or corrupted by noise. In this
;ase, image-processing techniques might be a useful tool for physicians to make their
lecision. With the advancement of computer technologies and digital image processing,
computer-based image processing system is also developed. In the current study, several
1pproaches for detection of cancerous lump in mammogram and ultrasound images are
ntroduced. These approaches based on contrast enhancement, filtering, segmentation
w'ith clustering techniques and pseudocoloring. There are 14 techniques have been
implemented under the category of contrast enhancement, 5 under filtering and 3 under
egmentation. This research emphasized on enhancing the image only on the interested
egion i.e.; on breast area, to reduce background influences. This technique is called
Embedded Enhancement Technique". Breast area is isolated from the background. The
10rder for breast area and background is revealed. Region growing method is used to
ind this border. Then other enhancement techniques stated before are applied into the
reast area. It can be observed from the experimental results that most of the
implemented algorithms can effectively enhance the contrast and sharpen the tested
images. Embedded techniques give better results than applying the enhancement
techniques on the whole images. A lot of details that could not be seen in the original
images have been clearly revealed. Th!s embedded technique was successfully beenĀ·
applied on mammogram images. It was then tried for ultrasound images. Unfortunately,
the results were not so good. To overcome this problem, the clustering techniques are
then applied on ultrasound images. A good result is obtained. These clustering
algorithms are suitable to be used on ultrasound images. Pseudocolouring technique also
~bows a good result. As other system, this system is designed to be user friendly,
suitable for those who are not familiar with computer. Therefore physicians could make
a good decision in determining direction of management treatment of the patient by
using this system.
Description
Keywords
Pemprosesan imej , Barah payudara