Development of detection system of kidney and non-kidney images for different ultrasound machines
The ultrasound machine is a well-known medical equipment being used widely to measure kidney size, shape, and position as well as assist the user in diagnosing kidney abnormalities such as stone, infection, and cysts. The interpretation of an ultrasound image totally depends on the operator’s kno...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/6461/1/24p%20JAUDAH%20ABD%20RANI.pdf http://eprints.uthm.edu.my/6461/2/JAUDAH%20ABD%20RANI%20COPYRIGHT%20DECLARATION.pdf http://eprints.uthm.edu.my/6461/3/JAUDAH%20ABD%20RANI%20WATERMARK.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The ultrasound machine is a well-known medical equipment being used widely to
measure kidney size, shape, and position as well as assist the user in diagnosing kidney
abnormalities such as stone, infection, and cysts. The interpretation of an ultrasound
image totally depends on the operator’s knowledge, skill and experience which may
be prone to misdiagnose. Previous researchers have built several detection systems to
detect kidney ultrasound images but does not consider the practical use of the detection
system in different ultrasound machines. This research proposes a detection system
which is able to detect the kidney ultrasound images regardless of the source of
ultrasound images. The detection system reduced the noise appeared in ultrasound
kidney image, enhanced the image and detect the kidney image regardless of the source
of the ultrasound kidney images. 188 kidney and non-kidney images were used in the
developing the system. The system was developed using MATLAB where the images
were enhanced using histogram equalization, filtered by Wiener filter, and segmented
manually by users using the mask tool. Then, texture analysis was done using
correlation in Gray Level Co-occurrence Matrix (GLCM) and classification was
performed using decision tree. Next, the detection system was developed by using
GUIDE MATLAB. In evaluating the effectiveness of the system, performance analysis
using confusion matrix was carried out. Performance analysis showed that the system
can detect kidney and non-kidney images from 4 different types of US machines with
a percentage accuracy of 72.97% while the sensitivity of the system is 81.97%. The
system specificity is 30.77%. The development of this detection system for ultrasound
kidney images has shown a promise and could be further improved for better
performances. This detection system hopefully could help the operator in getting a
second opinion on interpretation of ultrasound kidney images thus minimizing the
human error in misinterpretation of the images. |
|---|