Enhanced brain tumour detection and localization using ridgelet transform in MRI

The early detection and localization of brain tumours in magnetic resonance imaging (MRI) data play a pivotal role in diagnosis and therapy planning. Timely detection and accurate localization are crucial for effective diagnosis and therapy planning. The study aims to improve early detection and localization of brain tumours in MRI data by introducing a unified assessment system that integrates several image processing components, including localization, detection and classification. It also leverages the Ridgelet transforms unique capabilities for detection of precise the location of tumour. In contrast to other approaches, the study uses the Ridgelet transform as its sole approach for localizing tumours. This research incorporates the categorization of MRI into tumorous and non-tumorous and further into Glioma and meningioma by employing Ridgelet transform in conjunction with Grey Level Co-occurrence matrix for feature extraction and classification using Multi-Layer Preceptron (MLP). This further ensures thorough evaluation of the localization process. The Ridgelet-based methodology achieved an accuracy of 97.32% for classifying MRI into tumorous and non-tumorous classifications. Additionally, the study explores the use of the Radon transform in conjunction with GLCM for tumour classification, yielding results with an overall accuracy of 97.32%. The paper offers an innovative and efficient methodology and highlights the significant importance of the Ridgelet transform in tumour localization. The results are more robust when compared to known models, which highlights the potential contribution of Ridgelet transform to the advancement of brain tumour characterization in MRI data.