Hybrid Artificial Bee Colony and Adaptive Fuzzy K-Means for Detecting Abnormalities in Brain MR Images

A brain tumor is an abnormal mass of cells in the brain and can cause issues by pressing on the brain or infiltrating brain tissue. Not all brain tumors are cancerous; grades 1 and 2 are benign but may recur and become malignant. Symptoms vary depending on the affected area of the brain, as different regions control different functions. Medical image processing is the need for the investigation of 3D images of the humans' internal organs, typically collected from a Magnetic Resonance Imaging or Computed Tomography, to identify disorders, assist therapeutic procedures like incision preparation and in academic research. The multiparametric MR images of individuals with low- and high-grade gliomas that were manually segregated by skilled assessors constitute up the BraTS data set. Human separation of tumors is a laborious, time-consuming process that necessitates the delineation of sections by a human expert. Segmenting gliomas using autonomous algorithms for learning has been explored as a solution to these inadequacies. For glioma segmentation, the suggested method combines artificial bee colony (ABC) optimization and adaptive fuzzy k-means clustering (AFKM) that may be readily integrated into the medical procedure. ABC optimization efficiently finds optimal initial cluster centroids, overcoming Fuzzy K-Means' limitation of poor initialization and suboptimal results. ABC optimization balances global search and local refinement for efficient clustering, while Adaptive Fuzzy K-Means dynamically adjusts clusters, with ABC fine-tuning for optimal fit. The efficacy of the suggested method was validated with the help of metric parameters.