Abstract: A tumor (American English) or tumour (British English) is commonly used as a synonym for a neoplasm that appears enlarged in size. Neoplasm is an abnormal mass of tissue as a result of abnormal growth or division of cells. This tumor may be solid or fluid-filled. It can occur even in the brain. As the brain is well protected by the skull, early and in-depth detection techniques are needed for the identification of brain tumors which is one of the challenging tasks. Magnetic Resonance Imaging (MRI) technique is mainly used for analyzing the brain as the images produced are of high precision and applicability. Most of the tumour identification methods make use of different machine learning and segmentation techniques to provide improved detection accuracy. The challenge lies in accurate diagnosis in spite of improved existing techniques. The objective of the proposed research is to classify the brain MRI dataset for the existence or non existence of tumors. The proposed method uses Multilevel Local Ternary Pattern (MLTP) for pattern string generation. This is grouped for faster processing and for further classification using Support Vector Machine (SVM). The generation of pattern string gives the classification accuracy of 96% when compared to the existing classification techniques.
R. Meenakshi and P. Anandhakumar, 2014. An Improved Local Ternary Pattern Based Tumour Classification of MRI of Brain. Research Journal of Applied Sciences, 9: 99-103.