Enhancing Brain Tumour Multi-Classification Using Efficient-Net B0-Based Intelligent Diagnosis for Internet of Medical Things (IoMT) Applications

Abstract

Brain tumour disease develops due to abnormal cell proliferation. The early identification of brain tumours is vital for their effective treatment. Most currently available examination methods are laborious, require extensive manual instructions, and produce subpar findings. The EfficientNet-B0 architecture was used to diagnose brain tumours using magnetic resonance imaging (MRI). The fine-tuned EffeceintNet B0 model was proposed for the Internet of Medical Things (IoMT) environment. The fine-tuned EfficientNet-B0 architecture was employed to classify four different stages of brain tumours from the MRI images. The fine-tuned model showed 99% accuracy in the detection of four different classes of brain tumour detection (glioma, no tumour, meningioma, and pituitary). The proposed model performed very well in the detection of the pituitary class with a precision of 0.95, recall of 0.98, and F1 score of 0.96. The proposed model also performed very well in the detection of the no-tumour class with values of 0.99, 0.90, and 0.94 for precision, recall, and the F1 score, respectively. The precision, recall, and F1 scores for Glioma and Meningioma classes were also high. The proposed solution has several implications for enhancing clinical investigations of brain tumours.

Keywords

• brain tumour
• magnetic resonance imaging (MRI)
• multi-classification
• Internet of Medical Things (IoMT)
• EfficientNet