Human brain tumor classification and segmentation using CNN

The study of tumors in brain segmentation with classification through neuroimaging methodologies has become significant in recent years. A brain tumor, if not detected on time, maybe fatal. An improper tumor diagnosis might result in severe problems, as there are various tumors. Hence, the proper classification will help clinicians to provide an appropriate cure. Deep Learning may be a kind of an artificial intelligence that has recently achieved fantastic success in classification and segmentation tasks. This study uses a convolution neural network that classifies brain tumors using two public datasets, describing the different tumor forms (glioma, meningioma, and pituitary tumor) as with three glioma grades (as describes, Grade II, Grade III, and Grade IV). A public MRI imaging dataset includes 233 and 73 patients with 516 and 3064 images on T1-weighted images. Where methodology employs a 25-layer CNN model using T1-weighted Magnetic Resonance Imaging (MRI) images to evaluate our method's performance against previously published approaches in the field. Our method outperformed the other methods using the same dataset. The experimental results demonstrated that this proposed method achieved a tumor classification accuracy in study I is, 86.23.% using Adam optimizer, and study II is 81.6% using Sgdam optimizer. The proposed algorithm has produced impressive results in the classification and segmentation of MRI brain images. It will help clinicians to detect and classify brain tumors.