Classification of Meningioma MRI Images using First-Order Statistical Feature and Markov Random Field Modeling

Abstract

The brain is one of the largest organs in the human body that has a major role in the visual processing system, hearing, memory, body movement control, etc. By a risk of certain causes such as radiation or genetic disease, the meninges of the brain can experience a tumor called meningioma. 90% of these tumors are benign whose growth is very slow and physical symptoms appear slowly. If these symptoms indicate meningioma, doctors usually recommend doing magnetic resonance imaging (MRI). By scanning the brain, doctors can determine the location and size of meningioma. Processing and analysis of MRI images is an important part of MRI modalities, namely to extract MRI image information in an efficient and accurate way. The current development of image processing techniques in the medical field, allows for more in-depth analysis of MR images that have been used in hospitals. This study aims to analyze the texture and segmentation of MRI image of meningioma using the Markov Random Field (MRF) modeling method and first-order statistical feature extraction. Digital image samples obtained from open source radiopaedia.org. From the first part of discussion, it was shown that normal brain images have a range of mean values of pixel intensity in the area between sulcus and ventricles (dark gray); area of sulcus (light gray); ventricles (white) area of the brain respectively 51,32-87,91; 67,62-136,40; 224,26-254,98. The pixel intensity value of meningioma images in this study is above the image value of dark gray and below the image value of white.