Mining of Cancerous Region from Brain MRI Slices with Otsu's Function and DRLS Segmentation

摘要

Cancer is one of the deadly diseases and major causes of death in the human group. A significant number of methods are already proposed to detect and cure cancer in its early phase. Cancer cell growth in the brain section is known as the Brain-Cancer (BC) and the early identification and the cure may improve the life span of the patient. Magnetic-Resonance-Imaging (MRI) is normally considered to record the brain section and which can be considered for further checking. This article discusses a practice to extract and study the cancer segment from the MRIs of Flair, TIC, and T2 modalities. This paper implements Cuckoo-Search driven Otsu's thresholding to improve the MRI slice and Distance-Regularized-Level-Set (DRLS ) to extort the infected segment for further evaluation. The aim of this article is to discuss the merit of the modalities, such as Flair, TIC, and T2 in the examination of the MRI of BC. The result confirms that Flair modality helps to achieve precise tumor and edema mining compared to other modalities. Further, the average accuracy in Flair is found to be >91%.