Recent Advances in Multi-classification of Brain Tumour MRI Images using Deep Dynamic Parallel Convolutional Neural Network with Fully Termite Alate Optimization Algorithm: A Systematic Review

Abstract

Brain tumour classification using magnetic resonance imaging (MRI) has become a critical research area due to its importance in early diagnosis and treatment planning. Traditional diagnostic methods rely on manual interpretation, which is time-consuming and subject to variability. Recent advances in deep learning, particularly convolutional neural networks (CNNs), have significantly improved automated brain tumour classification accuracy. Studies show that deep learning models can achieve classification accuracy exceeding 97% for multi-class MRI datasets. Dynamic parallel CNN architectures further enhance performance by extracting multi-scale features through parallel convolutional pathways, improving robustness in handling tumour heterogeneity. Additionally, optimization techniques such as Particle Swarm Optimization (PSO), Genetic Algorithms (GA), and emerging bio-inspired methods like Termite Alate Optimization (TAO) have been employed to optimize hyperparameters and improve convergence. These optimization techniques enhance model performance and prevent local minima issues. This systematic review explores recent advancements in brain tumour MRI classification using deep learning and optimization techniques. It focuses on dynamic parallel CNN models integrated with optimization algorithms, highlighting their effectiveness in multi-class classification tasks. The study also discusses current challenges, including computational complexity, data scarcity, and model interpretability, and outlines future research directions for improving AI-based medical imaging systems.