A Novel and Robust Meningioma Tumor Identification using Modified Convolutional Neural Network

Brain tumors represent the most prevalent as well as violent kind of disease, with such a life expectancy that is rather low when they have reached their greatest serious stage. Therefore, the scheduling of therapy is an essential stage in the process of enhancing the sufferers' life satisfaction. This malignant type of the disease is distinguished by unchecked proliferation of brain cells within the skull, and it may be fatal. Because of the heterogeneity of tumor cells, classification of these cells is infamously challenging. Convolutional neural networks, also known as CNNs, are indeed the machine learning technique that is utilized the much more frequently for multimedia elements as well as the acknowledgement of brain tumors. A Dense EfficientNet (DEN) that uses min-max normalization was suggested as a classification method for meningioma tumors in this research. The recently designed system is a modification of EfficientNet that includes the addition of dense as well as drop-out layers. In such a comparable pattern, the study utilizes back propagation algorithm in order to heighten the comparison of the tumor cells. One of the advantages of utilizing a DEN model is that it would be able to more accurately classify a constrained image datasets. As a consequence of this, the approach that was suggested offers outstanding model effectiveness. According to the findings of the experiments, the suggested model had an accuracy of 98.15 percent even during learning phase and 97.89 percent even during testing phase. The recently developed DEN architecture has the potential to be a helpful decision-making weapon in the ongoing research into Meningioma diagnostic tests due to its high level of accuracy and favorable F1 score.