| Classification Of Tomographical Images Of Brain Tumors Using Artificial Neural Networks.
Basically, tomographical imaging is known to be an image formation method using projections. Recently, one of the techniques frequently used in imaging in medical diagnosis is tomographical imaging method. It is possible to see some tumors or other problems appearing in different organs of a body using tomographical imaging.
A tumor is formed when any cell can not be controlled. Since brain tumors are those growing inside the brain, the number of cells increases without control. Tumors are separated into two types: bening and malign. If the tumors are formed from the cells of the brain, they are called as primary brain tumors. The tumors developing in other systems of the body are called as metastatic brain tumors since they formed tumors by jumping to the brain (metastasize). For this reason different imaging methods are used for brain tumors. Mainly, the important ones are CBT (computerized brain tomography), MRI (magnetic resonance imaging), Pet Scan (positron emission tomography), DSA (angiography) and SPECT (single photon emission computed tomography).
In this study, the classification of different tumors formed in the brain for diagnosis. As the classifier, a technique, which is frequenctly used and known as an active classification technique, artificial neural network is used. The artificial neural network type used here is error back-propagation training algorithm. Using this classifier, a active classification is achieved by extracting feautre vectors of three different tumor images obtained from computerized brain tomographies belonging different patiens.
In the first part of this thesis, the tomographical imaging technique and methods are mentioned. In the second part, the artificial neural network (ANN) classifiers are mentioned. In the final part of the thesis, the method followed and ANN program used is introduced and the results are discussed.
Dostları ilə paylaş: |
