![]() | Only 14 pages are availabe for public view |
Abstract Magnetic resonance imaging (MRI) is one of the most widely used medical imaging technologies applied in many different medical procedures. Daily progress of medical information volume often leads to human errors in the manual analysis and elevates the need for automatic analysis. Thus, applying some tools to gather, classify, and evaluate medical data automatically is mandatory. Medical imaging issues are highly complicated owing to the importance of accurate diagnosis and disease treatment in healthcare systems. For these reasons, algorithms of automatic medical image analysis are utilised to help improve the accuracy and reliability of the medical images. AI methods like digital image processing and combinations with other techniques, such as fuzzy logic, machine learning, neural networks, and pattern recognition, are very important in the analysis and visualisation of medical images. The objective of this study is to investigate the purpose of artificial neural networks (ANN) in detecting bladder cancer at an early stage (for diagnosis), determining tumour stage (for prognosis), and examining the accuracy of MRI in T staging bladder cancer. Four algorithms are used for this purpose, namely Jordan/Eleman, Multilayer Perceptron (MLP), support vector machines (SVM), and Self-Organising Feature Map. A set of functional images taken through magnetic resonance (MR) was used. Results revealed that MLP yields better results than other algorithms. |