Numerical Simulation and Design of Efficient Brain Tumor Segmentation with Hybrid Fuzzy K-Means Clustering Technique

Abstract

The goal of this study is to detect brain cancers from MRI images using a Matlab GUI interface. Using the GUI, this application may employ numerous combinations of segmentation, filters, and other image processing methods to produce the best results. We begin by applying the Prewitt horizontal edge-emphasizing filter on the image. “Watershed pixels” are the next step in detecting tumours. The most essential aspect of this project is that all of the Matlab programmes function using a graphical user interface called “Matlab guide”. This allows us to experiment with different filter combinations and image processing approaches to get the optimal outcome for detecting brain cancers in their early stages. Scientists have been paying attention to the advancement of biomedical image processing; nonetheless, there have been certain issues with biomedical image processing, notably with MRI imaging. The one basic problem with the biomedical image segmentation (MRI image) method for image segmentation is that it varies considerably depending on the individual application. The segmentation software used with MRI imaging, for example, has different needs than the segmentation programme used with CT scan imaging. Furthermore, even if the images are from the same image programme, such as MRI, each image has its own quirks. These can differ from other MRI pictures, which will produce different results when read by the same segmentation tool. We suggested a unique MR brain image segmentation for recognising the tumour and locating the tumour area with enhanced accuracy and reduced processing time in this study. This dissertation examines a new hybrid clustering approach for speeding up computations and a binarization method for calculating area in mm2 using typography and digital imaging units. We compared the simulation results to existing techniques and the suggested shaft algorithm, after which we determined the tumour area and computed the CPU computing time. Finally, the suggested technique outperformed current algorithms while requiring less computational time.