Combining Multiple Style Transfer Networks and Transfer Learning For LGE-CMR Segmentation

Bo Fang; Junxin Chen; Wei Wang; Yicong Zhou

ICASSP 2022accepted0 citations

Combining Multiple Style Transfer Networks and Transfer Learning For LGE-CMR Segmentation

Bo Fang, Junxin Chen, Wei Wang, Yicong Zhou

Abstract

This paper presents an algorithm for segmenting late gadolinium enhancement cardiac magnetic resonance (LGE-CMR) in the absence of labeled training data. The proposed method includes a data augmentation part and a segmentation network. Multiple style transfer networks are employed for data augmentation to increase the data diversity, and then the synthetic images are used for training an improved U-Net. Finally, the trained model is fine-tuned with a few LGE images and labels. Experiment results demonstrate the effectiveness and advantages of the proposed method.

BibTeX

@inproceedings{icassp2022_combiningmultipl,
  title = {Combining Multiple Style Transfer Networks and Transfer Learning For LGE-CMR Segmentation},
  author = {Bo Fang and Junxin Chen and Wei Wang and Yicong Zhou},
  booktitle = {ICASSP 2022},
  year = {2022}
}