<![CDATA[Identification, segmentation and detection of brain tumor-infected parts on MRI images require precision and a long time. MRI of the brain has an important role, one of which is used for analysis or consideration before performing surgery. However, MRI images cannot provide optimal results when analyzed because of the presence of noise and the bone and tumor (clots of flesh) have the same appearance. Many studies related to brain tumor segmentation have been carried out before, and some of the good methods are CNN U-Net. We segmented brain tumors on MRI with U-Net. The purpose of this study was to analyze the results of changes in the number of neurons in the convolution layer of the U-Net architecture in segmenting brain tumors. We use two scenarios of changing the number of neurons at the U-Net convolution layer. The first scenario is the number of neurons successively at each level of the U-Net architecture [32,64,128,256,512], and the second scenario is [16,32,64,128,256]. And the results of scenario two can segment brain tumors on MRI images that resemble ground truth. The results of brain tumor segmentation in MRI images with the U-Net second scenarios have an average Dice value of 0.768.]]>
Copyrights © 2023
