A Novel h-CNN Architecture based Brain Tumor Classification

Abstract

This article presents an innovative approach to classifying brain ṭumors as ḅenign or ṃalignant using fused CT and MRI images. A novel hybrid-convolutional neural network (h-ĊNN) architecture is proposed, which leverages the complementary strengths of CT and MRI imaging modalities to enhance classification accuracy. The ĊNN architecture is designed to extract and integrate critical features from the fused images, providing a robust framework for ṭumor analysis. To further refine the classification process, a Vector Machine (ṢVM) is employed, enhancing the differentiation between ḅenign and ṃalignant ṭumors. The study demonstrates that combining ĊNN and ṢVM, called as hybrid CNN, significantly improves classification performance compared to traditional methods. Extensive experimentation on a comprehensive dataset of brain ṭumor images reveals the efficacy of the proposed approach, with results indicating superior accuracy, sensitivity, and specificity. This hybrid model not only advances the state-of-the-art in medical image analysis but also holds substantial potential for clinical application, offering a reliable tool for early and accurate brain ṭumor diagnosis. The integration of fused imaging techniques and advanced machine learning algorithms marks a significant step forward in the field of medical diagnostics, potentially improving patient outcomes through timely and precise intervention.