Unsupervised Medical Image Translation Using Cycle-MedGAN
This work addresses the challenge of medical image translation without paired data, which is often difficult to acquire in realistic scenarios, by providing a more effective unsupervised method for applications like PET-CT and MR imaging.
The authors tackled the problem of unsupervised medical image translation, where existing methods produce blurred or unrealistic images, by proposing Cycle-MedGAN with new non-adversarial cycle losses to minimize textural and perceptual discrepancies. The result showed improved performance in PET-CT translation and MR motion correction compared to other unsupervised approaches, as demonstrated through qualitative and quantitative comparisons.
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.