Image Generation combined with Super-Resolution: A Medical Image Pipeline

Cancer remains a leading cause of mortality worldwide, with breast and lung cancers accounting for many cases. Accurate and early diagnosis relies on high-quality medical imaging, yet data scarcity and limited resolution constrain robust computational tools. This work presents a pipeline integrating synthetic image generation with super-resolution to improve the quality and availability of medical imaging. Our approach splits the problem into two: anatomical coherence and variability are handled by 3D generative adversarial networks (GANs), while fine structural detail is enhanced via Real-ESRGAN. The framework was evaluated on breast MRI and lung CT datasets. Despite some limitations, results demonstrate that combining generative modelling with super-resolution can expand medical imaging datasets and improve fidelity, contributing to more precise and reliable cancer diagnosis, as confirmed by