Semi-automatic segmentation using MR images I.
Magnetic resonance imaging (MRI) is becoming clinically important in the assessment of joint injury and osteoarthritis because of its excellent soft tissue contrast. Segmentation of specific tissue structures is beneficial to the diagnosis and treatment of pathologies. A semi-automatic segmentation...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Final Year Project |
Language: | English |
Published: |
2009
|
Subjects: | |
Online Access: | http://hdl.handle.net/10356/16632 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
Summary: | Magnetic resonance imaging (MRI) is becoming clinically important in the assessment of joint injury and osteoarthritis because of its excellent soft tissue contrast. Segmentation of specific tissue structures is beneficial to the diagnosis and treatment of pathologies. A semi-automatic segmentation code based on a combination of histogrambased techniques, morphological operations and an active contour method was proposed to delineate the meniscus from knee MR images. The objective of the project is to
achieve a fast and accurate segmentation of the meniscus.
The code was originally designed using 1 set of MRI images, and subsequently tested on 3 other sets to validate its versatility to address different cases. The results were
compared to their manually segmented counterparts and computations of sensitivity were used as a gauge for accuracy. Implementation of the semi-automatic segmentation code on the original set generated an excellent mean sensitivity of 78%. However, the program was not as effective on the other MRI sets, giving less than 50% average for sensitivity. Nonetheless, segmentation of each image slice only took around 30 seconds.
In conclusion, a considerable amount of segmentation time was saved through the
use of the semi-automatic algorithm, but accuracy could still be improved for the other 3 sets so as to achieve adaptability which is a defining factor for a superior program. Future improvisations could consider taking preceding segmented images as a form of guidance when segmenting the subsequent ones. |
---|