MRI-to-CT Scan Conversion Using CycleGAN: A Deep Learning Approach

Radiological imaging has been a game changer for the diagnosis and treatment of various medical conditions, especially when paired with machine learning. However, the challenge of obtaining the paired dataset limits the innovation in this field. Obtaining MRI and CT scans for a single patient can be challenging due to practicality, availability, cost, and patient-specific limitations. This paper presents an approach to convert MRI scans into CT scans and vice versa using CycleGAN, which works on unpaired image-to-image dataset. The model is trained and evaluated using publicly available datasets and metrics such as structural similarity index measure (SSIM) and peak signal-to-noise ratio (PSNR) are used to assess the model’s performance. The proposed solution aims to enhance accessibility to medical imaging, reduce radiation exposure, and improve diagnostic efficiency.