Article Preview
Top1. Introduction
With the improvement of modern medical standards, medical imaging technology plays an increasingly important role in research on the medical diagnostic. Similarly, a human brain is centre of the nervous system (Choudhury et al. 2020; Anitha & Raja, 2018). A tumor of brain is a collection of uncontrolled increasing of these cells abnormally found in a different parts of the brain. However, brain tumor has become a key research topic in the medical field. For (Mittal & Kumar, 2019), the symptoms of a brain tumors may range from severe headaches and seizures to problems with vision and mental changes, depending on different parts of the body. In fact, early detection of tumor cells plays a major role in the treatment and recovery of patients. Diagnosing a brain tumor usually undergoes a very complicated and time-consuming process. In the medical imaging era, different medical imaging techniques namely X-ray, Magnetic Resonance Imaging (MRIs), Ultrasound, MRS (Magnetic Resonance Spectroscopy), and Computed Tomography (CT), have a great influence on the brain tumor detection and treatment process of patients (Jalali & Kaur, 2020; Amin et al. 2020). Furthermore, MRI (Magnetic Resonance Imaging) has demonstrated out as an effective instrument in the location of brain tumor with the assistance of MR Images (Kumar et al. 2017). Besides that, Magnetic Resonance Imaging (MRI) is the preferred way to diagnose a brain tumor, as it generates more detailed pictures than Computerized Tomography (CT) scans (Sarhan, 2020). MRI images help physicians study and diagnose diseases or tumors present in the brain (Ruba et al. 2020). This technique produces clear and high-quality images in various medical image formats. According to several researchers, a brain tumors can be classified into two types: benign and malignant. Benignant tumors have a homogeneous structure and don't contain disease cells while malign have a heterogeneous structure and contain malignancy cells (Jain & Godara, 2017).
Meanwhile, automated and accurate classification of MRI, X-ray, and CT brain images is extremely important for medical analysis and interpretation. Over the last few years, numerous techniques are devised for brain tumor classification using different deep learning methodologies considering imaging modalities like MRI, CT, and so on. According to (Deepak & Ameer, 2019), the classification of tumors using brain images is a challenging task due to two issues. The first issue is that the brain tumor pose high variations in contrast to size, intensity, and shape. The second issue is that the tumors poses many pathological types, which pose the same manifestation. Moreover, brain tumor classification is an important problem in computer-assisted surgery systems, healthcare artificial intelligence systems in order to give yield more advantages for the medical diagnosis. Apart from this, Deep Learning (DL) has also been able to obtain significant results in the domain of medical diagnosis. Deep Learning has become one of the most common techniques that have achieved better performance in many areas, especially in medical image analysis and classification (Al ayoubi et al. 2020). According to (Ali Khan et al. 2020; El Kader Isselmou et al. 2019), because of Deep Learning, significant advancement has been made in medical science like the medical image processing technique which helps doctors to diagnose the disease early and easily. For (Kaur & Singh, 2020), Deep learning (DL) techniques are widely used in the automatic analysis of radiological images. Likewise, Deep Convolutional Neural Networks (DCNNs) has been used in the medical imaging classification and grading since it does not require preprocessing or features extraction before the training process (Sasikala & Kumaravel, 2008). For (Alqudah et al. 2019), DCNNs are designed to minimize or canceling in sometimes the data pre-processing steps and usually are used to deal with raw images. Deep Convolutional Neural Networks (DCNNs) models is a class of deep, feed-forward artificial neural networks that has successfully been applied to analyzing visual imagery (Zabalza et al. 2016). DCNNs use relatively minimal pre-processing compared to other image classification algorithms.