Brain Tumor Detection from MRI Images by K-Means and Level Set Method

Abstract

Tumor is the preeminent explanation of fatality in monetarily developing and developed nations and due to tumor second highest fatality occurred around the world. In disease, cells divide and expentially grow, shaping injurious tumors, and strike close body parts. To detect brain tumour various techniques available and out of them image segmentation is one of them. In last few years various image segmentation algorithms used to analyze MRI images to detect brain tumour. But there are various challenges exist in detecting brain tumour like compression, precise detection. There are numerous techniques which are used to detect brain tumor for example SVM, K-means, Fuzzy C-means, Genetic algorithm, Image processing and many more. In overcoming this challenge, in our research an integration approach is implemented using improved K means and level set technique. The result of the comparative analysis shows that implemented hybrid algorithm is better than existed algorithm in term of accuracy, standard deviation, area, mean and perimeter. Proposed algorithm shows accuracy around 80% wheras reference work has accuracy only around 55 %. In future work more robust algorithms based upon machine learning can be implemented to enhance accuracy.