ShQDFHNet: Shepard quantum dilated forward harmonic net for brain tumour detection using MRI image.

Introduction: One of today's major health threats is brain tumours, yet current systems focus mainly on diagnostic methods and medical imaging to understand them. Here, the Shepard Quantum Dilated Forward Harmonic Net (ShQDFHNet) is developed for brain tumour detection using MRI scans.

Methods: It starts by enhancing images with high boost filtering to highlight key features, then uses Log-Cosh Point-Wise Pyramid Attention Network (Log-Cosh PPANet) for accurate tumour segmentation, guided by a refined Log-Cosh Dice Loss. To capture texture details, features like Spatial Grey-Level Dependence Matrix (SGLDM) and Gray-Level Co-occurrence Matrix (GLCM) are extracted. The final detection uses ShQDFHNet, combining Shepard Convolutional Neural Network (ShCNN) and Quantum Dilated Convolutional Neural Network (QDCNN), with layers enhanced by a Forward Harmonic Analysis Network.

Results: ShQDFHNet achieved strong performance on the Brain Tumour MRI dataset, with 90.69% accuracy, 91.14% True Positive Rate (TPR), and 90.61% True Negative Rate (TNR) using K-fold of 9.

Discussion: The use of high boost filtering, Log-Cosh PPANet, and texture-based features improves the input data quality and enables accurate tumor segmentation in MRI scans. The proposed ShQDFHNet model improves feature learning and achieves strong performance on brain tumor MRI data.