Efficient Classification of Prostate Cancer Using Artificial Intelligence Techniques

• Published in: SN computer science Vol. 5; no. 4; p. 389
• Main Author: Baazeem, Rami Mohamdfowzi Yaslam
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 02.04.2024; Springer Nature B.V
• Subjects: Algorithms; Artificial intelligence; Bladder; Breast cancer; Cancer therapies; Classification; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data Structures and Information Theory; Diagnosis; Diagnostic systems; Discriminant analysis; Feature extraction; Hormones; Image quality; Image segmentation; Information Systems and Communication Service; Life expectancy; Machine Intelligence and Smart Systems; Machine learning; Magnetic resonance imaging; Medical imaging; Mortality; Multilayer perceptrons; Noise reduction; Original Research; Particle swarm optimization; Pattern Recognition and Graphics; Penis; Prostate cancer; Signal to noise ratio; Similarity; Software Engineering/Programming and Operating Systems; Sperm; Support vector machines; Tumors; Vision; Asia; Magnetic resonance imaging (MRI); Support vector machine (SVM); Multilayer perceptron (MLP); Linear discriminant analysis (LDA)
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-024-02745-0

As the primary cause of illness and death for men, Prostate Cancer (PCa) is a serious worldwide health problem. To maximise treatment results and raise patient survival rates, prostate cancer diagnosis and timing are essential. The rising death rate of PCa could be reduced with better treatment planning and diagnosis, both of which depend on early and accurate detection. In this paper, a comprehensive computer-aided diagnostic (CAD) system that transforms prostate cancer diagnosis is introduced. The system includes crucial steps like feature extraction, feature classification, prostate segmentation, and image denoising, providing an all-encompassing strategy for early and precise detection. A careful evaluation of the denoising filters’ effectiveness on 300 mp-Magnetic Resonance Imaging images focuses on preserving key structural features. Notably, anisotropic and Non-Local Means filters are shown to be the best options for reducing noise while maintaining image quality. The quality of diagnostic images is improved as a result of the standardization of denoising techniques in medical imaging. A remarkable 7.8% improvement in accuracy is achieved by the study’s integration of Particle Swarm Optimization (PSO) into the segmentation process. The ability of PSO-based segmentation algorithms to manage a variety of data sources and medical imaging modalities effectively is highlighted by the flexibility and robustness of these algorithms. The study uses support vector machine, multilayer perceptron, and linear discriminant analysis machine learning algorithms to accurately classify benign and malignant prostate cancer cases. A thorough evaluation framework that includes a variety of assessment metrics, including the Peak Signal Noise Ratio, Structural Similarity Index, Feature Similarity Index, Mean Opinion Score, and a Final Score, guarantees a thorough examination of the performance of the suggested method.