Brain Tumour Segmentation in FLAIR MRI Using Sliding Window Texture Feature Extraction Followed by Fuzzy C-Means Clustering

2. Background

Medical image processing, via MRI and other means (e.g., CT Scan), plays a key role in the analysis of brain tumors; in turn, this process supports how the tumor and its treatment plan may best be handled. Primarily used in the diagnosis of brain tumor and for treatment planning, the MRI is a powerful non-invasive medical imaging modality for brain scans (“MRI Basics”, 2018). It offers a variety of valuable features, including multiplanar capabilities as well as prospective of tissue characterization with no teeth and bone artefacts. Various MR sequences are generated by changing the excitation times during the acquisition of an image (e.g., Clark, Hall & Goldgof, 1998; Işın, Direkoğlu & Şah, 2016). These various MRI sequences produce diverse kinds of tissue contrast images, creating very significant basic information to allow the proficient extraction of tumors alongside their sub-regions (e.g., El-Dahshan, Mohsen, Revett & Salem, 2014; Işın et al., 2016; Saxena, Garg & Pattnaik, 2019).

Figure 1 portrays the three different MRI sequencing standards for brain tumor diagnosis known respectively as the T1, T2 and FLAIR MRI Sequences (Ahmed, Iftekharuddin &Vossough, 2011).

Figure 1.

Types of MR Sequences