Thesis

It is an iterative and time-consuming process to classify, segment, and detect the area of infection in MRI images of brain tumors. A theory of image processing can imagine the various anatomical structure of the human body. It is difficult to visualize the structure of the abnormal human brain with simple imaging techniques. The magnetic resonance imaging method Differentiates and clarifies the neuronal architecture of the human brain. The MRI technique includes several imaging methods that analyze and capture the internal structures of the human brain. In this study, we focused on noise reduction techniques, grey-level matrix (GLCM) extraction features, segmentation of DWT-based brain tumor areas that reduce complexity and improve performance. This, in turn, eliminates the noise that can be generated after segmentation by morphological filtering. Probabilistic neural network classifiers were used in brain MRI images to check tumor location and check output accuracy. The classifier was used in MRI imaging of the brain to test for tumor location and to test output accuracy. Experimental results achieved nearly 100% accuracy in identifying normal and abnormal tissues from brain MRI images, which demonstrates the effectiveness of the proposed technique.