COMBINATION OF GABOR FILTER AND MAX OPERATOR IN BRAIN ABNORMALITY DETECTION MODEL
Brain abnormality always becomes a crucial issue in the medical field. Early procedures which are generally conducted in dealing with the problem are medical <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/17278 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Brain abnormality always becomes a crucial issue in the medical field. Early procedures which are generally conducted in dealing with the problem are medical <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
image aquisition and the interpretation. Problems emerges when the medical image cannot be easily interpreted because of unexplicit appearance of the abnormality in the image. This research aimed to contructs methods and parameter determination of segmentation process and recognition system for MRI image of brain with abnormality. Recognition system which was used based on the combination of Gabor filter and MAX operator as similar methods to the human visual system. In both methods, there are two parameters which the determinations are crucial to determine the object recognition ability, especially in the MRI image recognition. Those parameters are scale factor and RF (receptive field) size. First, segmentation process was conducted using thresholding method, which was then continued by filtering process to recognize the non-abnormal area. The connected-component labeling method was also employed, both for early and advance detection process of abnormality. From the segmentation result, the MRI <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
images in the same dataset were able to be separated between those are normal and abnormal brain images. The result of images separation was then used as the input of object recognition system with the modification in the factor scale and RF size. At the final result, it was found that the best scale factor and RF size for MRI image recognition were different with the real image recognition. The best scale factor and RF size for MRI images are 21/4 and 3x3, respectively. The constructed segmentation method is also proven well in isolate and localize abnormality in the brain MRI images. |
|---|