Affiliations:
Department of Medical Laboratory Technology (MLT), Faculty of Applied Medical Sciences, King Abdulaziz University, Rabigh, Saudi Arabia
Prostate cancer is one of the most common and lethal cancers among men worldwide, often developing without symptoms or with only minor signs, which makes early detection difficult. This study introduces a novel deep learning framework that integrates a Conditional Generative Adversarial Network (CGAN) with a weighted ensemble model consisting of ResNet (15%), Xception (50%), and Inception (35%). The model was trained and evaluated on the publicly available PROSTATEx image dataset, with performance measured using accuracy, precision, recall, and F1-score. The proposed weighted ensemble achieved 98.25% accuracy, 96.75% precision, 97.58% recall, and 96.65% F1-score, outperforming the soft voting ensemble, which obtained 95% accuracy, 92% precision, 94% recall, and 93% F1-score. Comparative analysis with state-of-the-art methods and cross-validation further confirmed the robustness of the proposed approach. These findings suggest that integrating CGAN with a weighted ensemble significantly enhances prostate cancer detection and holds strong potential for clinical application in early diagnosis and patient management.
Prostate cancer, Deep learning, Generative adversarial network, Ensemble model, Medical imaging
https://doi.org/10.21833/ijaas.2025.12.008
Munshi, R. M. (2025). An automated approach for prostate cancer detection using CGAN data augmentation with a weighted ensemble of transfer learning models. International Journal of Advanced and Applied Sciences, 12(12), 75–86. https://doi.org/10.21833/ijaas.2025.12.008