An Approach for Classification of Brain Tumor using Fully Connected Deep Convolutional Neural Network

Convolutional Neural Network (CNN) structures have played a vital part in various medical sectors due to their efficiency. It is likely becoming the most significant early identification approach for malignant tumor masses as digital systems and CNN technology advance. The research focuses on the early diagnosis and detection of tumors through deep learning with convolutional neural networks and convolution neural network approaches. It focuses on the processing, fragmentation, and categorization of various brain tumor pictures into malignant or benign tumors. The evaluation is done by incorporating several layers of CNN using a 70:30 and 80:20 splitting ratio. Conventional deep learning classifiers perform the comparative analysis of the routine of projected CNN prototypical, and the planned CNN model perceived accuracy of 97.92% using VGG16 and 98.20% using VGG19 deep learning frameworks. Furthermore, it is observed from experimental findings that the performance of various pre-trained models of VGG19 performs better than that of other models.