An Efficient VGG-RIME-ELM Model for Brain Tumor Detection
Hui Xu, Rui Zhong, Zhongmin Wang, Limin Xu, X Y Wang · 2024
Brain tumor detection is a critical area in medical diagnostics, where accurate and timely detection can significantly improve patient outcomes. This paper proposes an efficient hybrid model for brain tumor detection named VGG-RIME-ELM, which integrates a pre-trained VGGNet, the RIME optimization algorithm, and the Extreme Learning Machine (ELM). The proposed VGG-RIME-ELM model utilizes the feature extraction capability of VGG, and the selected features in the dataset are fed to the ELM. To further improve the performance and accuracy of ELM, we incorporate the RIME optimization algorithm, an effective metaheuristic algorithm, to search for optimal configurations of ELM. To validate our proposed VGG-RIME-ELM model, we conduct numerical experiments in the public brain tumor detection dataset from Kaggle. Five baseline deep learning models are employed for comparison. Comprehensive experimental results confirm the competitiveness of our proposal against baseline algorithms, which have broad perspectives for real-time clinical applications.