Unsupervised Medical Image Translation Using Cycle-MedGAN

Abstract

Image-to-image translation is a new field in computer vision with multiple potential applications in the medical domain. However, for supervised image translation frameworks, co-registered datasets, paired in a pixel-wise sense, are required. This is often difficult to acquire in realistic medical scenarios. On the other hand, unsupervised translation frameworks often result in blurred translated images with unrealistic details. In this work, we propose a new unsupervised translation framework which is titled Cycle-MedGAN. The proposed framework utilizes new non-adversarial cycle losses which direct the framework to minimize the textural and perceptual discrepancies in the translated images. Qualitative and quantitative comparisons against other unsupervised translation approaches demonstrate the performance of the proposed framework for PET-CT translation and MR motion correction.

Submitted to EUSIPCO 2019, 5 pages

References in corpus (11)

• Generative Adversarial Networks
• Efficient Multi-Scale 3D CNN with Fully Connected CRF for Accurate Brain Lesion Segmentation
• Generative Adversarial Network in Medical Imaging: A Review
• Compressed Sensing MRI Reconstruction using a Generative Adversarial Network with a Cyclic Loss
• Perceptual Adversarial Networks for Image-to-Image Transformation
• Cycle Consistent Adversarial Denoising Network for Multiphase Coronary CT Angiography
• Generative Adversarial Networks for Image-to-Image Translation on Multi-Contrast MR Images - A Comparison of CycleGAN and UNIT
• Adversarial Inpainting of Medical Image Modalities
• Retrospective correction of Rigid and Non-Rigid MR motion artifacts using GANs
• Stacked U-Nets: A No-Frills Approach to Natural Image Segmentation
• Deep CT to MR Synthesis using Paired and Unpaired Data

Cited by in corpus (16)

• Anomaly Detection in Medical Imaging -- A Mini Review
• Age-Net: An MRI-Based Iterative Framework for Brain Biological Age Estimation
• CyTran: A Cycle-Consistent Transformer with Multi-Level Consistency for Non-Contrast to Contrast CT Translation
• Normalization of breast MRIs using Cycle-Consistent Generative Adversarial Networks
• Unsupervised Adversarial Correction of Rigid MR Motion Artifacts
• AeGAN: Time-Frequency Speech Denoising via Generative Adversarial Networks
• Is Image-to-Image Translation the Panacea for Multimodal Image Registration? A Comparative Study
• Bridging the gap between paired and unpaired medical image translation
• Deep Learning Techniques for Inverse Problems in Imaging
• Unsupervised MR Motion Artifact Deep Learning using Outlier-Rejecting Bootstrap Aggregation
• Adversarial Uni- and Multi-modal Stream Networks for Multimodal Image Registration
• Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning
• A Strictly Bounded Deep Network for Unpaired Cyclic Translation of Medical Images
• ASL to PET Translation by a Semi-supervised Residual-based Attention-guided Convolutional Neural Network
• Multi-Texture GAN: Exploring the Multi-Scale Texture Translation for Brain MR Images
• Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET