Beyin tümörü MR görüntülerinin havza eşiği kullanarak segmentasyonu

The detection of brain tumors using magnetic resonance imaging (MRI) brain scans has become one of the most active areas of research in the field of medical image processing. The primary purpose of the system is the detection of the tumor. In biomedical imaging, detection plays a pivotal function. In this study, MRI brain images are used to detect tumors. If a clear image of the brain cannot be obtained and the location, size, and type of the tumor are known, the patient will be administered incorrect doses, resulting in the destruction of healthy brain tissue. Due to the dearth of radiology-trained physicians, which lengthens the time required to make a diagnosis due to the doctors' workload, the algorithm's ability to reduce the time and speed required to make a diagnosis makes it extremely valuable if implemented. In this study we propose a new method for segmenting the tumor territory on MRI scans in order to locate and isolate it. Initially, the watershed approach is used to the images, followed by a second segmentation to isolate the brain tumor. MATLAB and Maya will be utilized for all application and division activities. After including a fitness function into the algorithm, the DICE- and JACCARD-segmented images will be scored to determine how well they compare to one another. When a 0.1 mm threshold was applied to either process, the resulting accuracy was outstanding. When comparing the results of the 3D segmentation to those of other models using interscan evaluation, discrepancies more than 0.1 mm were observed in 16 of 21 instances. If other brain tumor segmentation algorithms use the standard threshold for segmentation, they will miss areas of the tumor. By modifying the segmentation threshold, we were able to recreate previously nonexistent brain regions and assess how these alterations affected the image quality. Both the Jaccard and Dice segmentation scores hit an all-time high of 0.95 after the threshold change.