Capsule GAN for prostate MRI super-resolution

Prostate cancer is a prevalent disease among adult men. One in seven Canadian men diagnosed with this their lifetime. Super-Resolution (SR) can facilitate early diagnosis and potentially save many lives. In paper, robust accurate model proposed for prostate MRI SR. For the first time, MSG-GAN CapsGAN are utilized simultaneously high-scale medical The trained on Prostate-Diagnosis PROSTATEx datasets. outperformed state-of-the-art SR all similarity metrics substantial margins. $$8 \times$$ SR, 19.77, 0.60, 0.79 achieved Peak Signal-to-Noise Ratio (PSNR), Structural SIMilarity index metric (SSIM), Multi-Scale (MS-SSIM), respectively. A new task-specific assessment introduced as well. classifier severe detection. drop accuracy of when dealing super-resolved images used to evaluate ability detail reconstruction models. surpassed work 6% margin. also more compact comparison related architecture has 45% less number trainable parameters. step towards an efficient general platform.