Abstract
Chronic Stress is one of the impacts of tumor formation in the human body. Radiation exposure and some family background of brain tumors are also harmful factors for tumor formation. There is no such correct explanation of what risk factors the primary tumor formation is possible. The Brain tumor is the most crucial and painful disease in the current generation. People even face difficulty in detecting the risk factors of the brain tumor. When people find a brain tumor at a starting stage, and then treating the tumor will be much easier. Before treating the tumor, they must know about every detail of the tumor, like the tumor's size, shape, and soon. They have three different ways: Machine Learning, Deep learning, and computer vision techniques. Image segmentation in computer vision is the method chosen in our proposed model. The image segmentation separates the tumor tissues from the normal tissues. A combination of Region-Growing Segmentation and threshold segmentation is used in the proposed model. Firstly, the input is taken as an MRI image of a brain tumor, and for the pre-processing, the wavelet-transform and an unsharp filter are used. Those images are used for image segmentation using the combination of region-growing Segmentation and threshold segmentation. Now, finally, do classification using CNN. Finding the accuracy, F1 score, precision, and recall for the ranking of the model. This proves that best outcome is achieved for MRI image segmentation of brain tumors.
Keyword
: Region-growing, Threshold, Wavelet transform, CNN, Unsharp filter.
PDF Download (click here)
|