Automated brain tumor recognition using equilibrium optimizer with deep learning approach on MRI images

Abstract

Brain tumours (BT) affect human health owing to their location. Artificial intelligence (AI) is intended to assist in diagnosing and treating complex diseases by combining technologies like deep learning (DL), big data analytics, and machine learning (ML). AI can identify and categorize tumours by analyzing brain imaging approaches like Magnetic Resonance Imaging (MRI). The medical sector has been promptly shifted by evolving technology, and an essential element of these transformations is AI technology. AI model determines tumours’ class, size, aggressiveness, and location. This assists medical doctors in making more exact diagnoses and treatment plans and helps patients better understand their health. Also, AI is used to track the progress of patients through treatment. AI-based analytics is used to predict potential tumour recurrence and assess treatment response. This study presents Brain Tumor Recognition using an Equilibrium Optimizer with a Deep Learning Approach (BTR-EODLA) technique for MRI images. The BTR-EODLA technique intends to recognize whether or not a BT presence exists. In the BTR-EODLA technique, median filtering (MF) is deployed to eliminate the noise in the input MRI. Besides, the squeeze-excitation ResNet (SE-ResNet50) model is applied to derive feature vectors, and its parameters are fine-tuned by the design of the EO model. The BTR-EODLA technique utilizes the stacked autoencoder (SAE) model for BT detection. A sequence of experiments is performed to ensure the improved performance of the BTR-EODLA technique. The investigational validation of the BTR-EODLA technique portrayed a superior accuracy value of 98.78% over existing models.