Brain Tumor Detection Using Extended Optimization Based Multi-Layer Perceptron

Classifying brain tumors is indispensable for lowering death rates and raising patient survival rates. It is essential to effectively comprehend brain tumors in order to evaluate the tumors and support patients take suitable treatment based on their classifications. There are several imaging methods for brain tumor identification, comprising Magnetic Resonance Imaging (MRI), which is commonly employed owing to its image quality and the fact that it relies on non-ionizing radiation. The current models that use these images have a number of drawbacks, including limited efficiency, high parameters, and poor accuracy. In this paper, an effective brain tumor detection system that detect tumors is built using an improved optimization based machine learning (ImpOML) model. The proposed model uses Extended Sea-Horse assisted Multi-Layer Perceptron (ExSH-MLP) model to categorize cancers and healthy brain tissue. To improve the training of MLP, an enhanced sea-horse optimization (ESHO) approach is employed to optimize the MLP neural network. For experimentation, the Brats-2021-task1 dataset is used mainly to show the value of the ImpOML model. This work provided a comparative examination of recent methods to demonstrate the efficacy of the ImpOML approach. The results showed that the ImpOML model has an effective accuracy of 98.5%, which is higher than the existing methods.