Adopting Run Length Features to Detect and Recognize Brain Tumor in Magnetic Resonance Images

Abstract

Early detection of brain cancer considered vital and attracted attention. In this study, a designed software presented to detect and recognise brain tumours. The segmentation based threshold used to detect the region of interest in Magnetic Resonance Imaging (MRI) images. Texture features were extracted using grey level run length matrix (GRLM), then detect tumours in MRI image and features image using a segmentation-based threshold technique. Location of the tumour in MRI and its features determined using the histogram as well as behaviour complement images for each feature. The geometrical characteristics determined of the tumour image as well as the complement image such as size, location, area and dimensions. The detection results depending on the segmentation technique were very effective in the separation of the entire tumour area. The quality of the texture features using GRLM has high accuracy by means separation of the tumour area from the complement area.