Continuous Conversion of CT Kernel Using Switchable CycleGAN With AdaIN

doi:10.1109/TMI.2021.3077615

Return

Continuous Conversion of CT Kernel Using Switchable CycleGAN With AdaIN

2021-11-01

Eung Yeop Kim

Jong Chul Ye *

DOI：10.1109/TMI.2021.3077615

Original

Save

View PDF

Abstract

En 中文

X-ray computed tomography (CT) uses different filter kernels to highlight different structures. Since the raw sinogram data is usually removed after the reconstruction, in case there is additional need for other types of kernel images that were not previously generated, the patient may need to be scanned again. Accordingly, there exists increasing demand for post-hoc image domain conversion from one kernel to another without sacrificing the image quality. In this paper, we propose a novel unsupervised continuous kernel conversion method using cycle-consistent generative adversarial network (cycleGAN) with adaptive instance normalization (AdaIN). Even without paired training data, not only can our network translate the images between two different kernels, but it can also convert images along the interpolation path between the two kernel domains. We also show that the quality of generated images can be further improved if intermediate kernel domain images are available. Experimental results confirm that our method not only enables accurate kernel conversion that is comparable to supervised learning methods, but also generates intermediate kernel images in the unseen domain that are useful for hypopharyngeal cancer diagnosis.

Keywords:

Kernel

Computed tomography

Generators

Switches

Image reconstruction

Interpolation

Deep learning

Computed tomography

reconstruction kernels

cycle-consistent adversarial networks

style transfer

adaptive instance normalization (AdaIN)

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.