Article Preview
Top1. Introduction
The brain tumor (Alsabti et al., 1997; Amarapur, 2020) is a collection of abnormal cells in human brain. The human skull is a very rigid part, enclosed by the brain. Any type of unwanted growth inside the skull can create the problem, as it is the restricted area. The brain tumor can be malignant (cancerous) or benign (non-cancerous). The malignant tumor growth increases the pressure in the skull, and enhances the probabilities to damage the brain. The brain tumor is one of the life threatening. The human brain tumors are classified as primary tumor or secondary tumor. Most of the primary brain tumor or benign, originates in human brain. The secondary class of tumor is caused by any other cancer cells in the organ, such as breast, lung, skin, kidney or chest. These cells spread in the human brain from these organs, reason of metastatic brain tumor.
The primary brain tumors generally develop from human brain cells, glands, nerve cells or membrane that covers our brain called meninges. The adult people have a common type of brain tumor called gliomas and meningiomas (Amin et al., 2020). The gliomas is developed inside glial cells. Gliomas is the most aggressive and common tumor that can lead the short span of life in its higher grade. To detect the tumor in early stage, and proceed further for the treatment, it is essentially required to consult from oncologist or radiation oncologist. It becomes very difficult to survive for the patient in the gliomas tumor.
The treatment includes surgery, radiotherapy, and chemotherapy or a combination of all these, based on the necessities of the patient. The patient hardly survives for two years in such cases. The common diagnosis methods are Computed Tomography (CT), Magnetic Resonance Imaging (MRI) (Bisht & Kumar, 2019; Chahal et al., 2020) and Position Emission Tomography (PET) Scan. MRI is very popular among all these methods for successful treatment and diagnosis of the brain tumor. If the patient has MRI of his head, a special due helps the doctor to detect the tumor. The CT scan is based on the radiations, but MRI is different from the CT scan. It provides the detailed picture of brain that can be used for further analysis and diagnosis. The segmentation of the gliomas tumor is very essential for the treatment of the patient and taking follow-ups. To characterize the tumor, manual segmentation can be done, but it is a time-consuming process and causes errors. Therefore, automatic brain tumor image segmentation and detection is required. There are many techniques to segment the brain tumor, such as Partial Differential Equation (PDE) based segmentation method (Gonzalez et al., 2002; Soille, 2013), Artificial Neural Network (ANN) based segmentation. The general methods of image segmentation are thresholding, edge based segmentation, region based segmentation, clustering based segmentation, watershed, ANN and PDE method (Saritha & Amutha Prabha, 2016).
The problem statement of the research work is to study the difficult image segmentation methods such as level set method, Watershed algorithm, Otsu’s method, K- Means Clustering, Discrete Wavelet Transform (DWT) and evaluate the performance of these segmentation methods based on parameters recall, precision, accuracy and MATLAB response time. The Machine-learning algorithms will help to estimate the accuracy of these algorithms, used for brain tumor MRI image segmentation and detection application.