Analysis of Magnetic Resonance Imaging of Brain Tumour using Fuzzy Set Rules

Abstract

Introduction: Cancer becomes life threatening once the expansion of tissues in human brain turns into an uncontrolled growth. In detection of brain tumours, Magnetic Resonance Imaging (MRI) images give better results when compared to Computerised Tomography (CT) scan and X-ray. Malignant tumours can be detected with the help of image processing and machine learning techniques. These techniques detect even a small abnormality in the human brain following a four-stage process which includes preprocessing, segmentation, feature extraction and optimisation. Aim: To predict a brain tumour using Fuzzy minimum-maximum rule in MRI. Materials and Methods: The medical challenge is how can we rapidly and precisely diagnose brain lesions. It is difficult when using standard image analysis to differentiate the benign and malignant lesions. Hereby, authors have used a Fuzzy imaging algorithm in data set of 253 brain tumour images of high grade tumour colllected from kaggle.com.This paper proposes a fuzzy min-max image processing algorithm. Image processing includes four stages- preprocessing, segmentation, feature extraction and accuracy detection. Brain tumours are located using various algorithms at each of these stages. Results: There were more than 20 features which can be taken into consideration when using the Fuzzy image algorithm. The proposed method in its current form achieves an accuracy of about 95% by considering seven features. Conclusion: The study revealed that age, shape, contour, blood supply, capsule of tumour, oedema, post contrast enhancement, cyst generation signal intensity of T-1 weighted image etc. were the important investigation parameters. Location and size are important to the domain experts, but due to their complexity these parameters are not considered here.

Keywords

• brain lesions
• fuzzy algorithm
• image processing
• oedema
• segmentation methods