Abstract An automated neurological disorder detection system which is intended to localize brain tumor using computer vision on Magnetic Resonance Imaging (MRI). The most widespread and aggressive brain tumors are gliomas, at their highest point, they lead to a much more limited life span. Therapy preparation is also a crucial step in ensuring a better quality of life for patients with oncology. Magnetic resonance imaging (MRI), It is used for inspecting the mechanisms & components of the human body also for medical diagnosis, to find the stage of the disease and to follow up without ionizing radiation exposure. The substantial spatial and structural changeability of brain tumors makes the process of segmentation more difficult. Consequently, depending on Convolutional Neural Networks (CNN), an automated and consistent segmentation approach is used. Through use of small kernels facilitates the construction of a deeper architecture, Owing to the fairly low number of network weights, with such a positive impact on overfitting. It also investigates the use of normalization of intensity as a pre-processing phase, which is not ubiquitous in segmentation techniques based on the Convolution Neural Network, but it has successfully demonstrated effectiveness in segmenting brain tumors in Magnetic Resonance Imaging (MRI) together with data augmentation.
Pranav Shetty, Suraj Singh, Rasvi Jambhulkar, Kajal Sheth, Deepali Ujalambkar. A survey on detection of brain tumor using computer vision and machine learning technique. Int J Appl Res 2021;7(3):84-88. DOI: 10.22271/allresearch.2021.v7.i3b.8365