E-DCNNBAE: design of a novel segmentation and classification approach for brain tumour prediction.

Context: Health plays an essential role in human life, particularly brain health, which supports vital system functions. Magnetic Resonance Imaging (MRI) is a crucial tool for diagnosing brain-related disorders and provides large datasets suitable for artificial intelligence, especially image classification. Objective: The aim of this study is to develop an accurate and reliable deep learning-based framework for the automatic classification of brain tumours (glioma, meningioma, and pituitary tumours) from MRI images. Materials and methods: Pre-processing is performed using the Adaptive Contrast Enhancement Algorithm (ACEA) and a median filter. For segmentation, Weighted Fuzzy C-means clustering (WFCMC) ensures reliable pixel assessment, normalisation, and generalisation. Image regularisation is achieved by scaling and analysing data loss to maintain pixel stability. A novel Ensembling Dense Convolutional Neural Network with Bayesian Auto-encoder (E-DCNNBAE) model is proposed for classification. Results: The performance of model is evaluated using accuracy, precision, recall, F-score, and AUC. It achieves 97% accuracy, 99% precision, 98% recall, 98% F-score, and 98% AUC. Conclusion: This study improving early diagnosis and treatment of brain diseases.