TranMix: An improved Mixup for Medical Image Segmentation
Deep learning is powerful for medical image segmentation; but it still faces the challenge of limited training data. To alleviate the challenge, we propose TranMix, a hybrid approach that combines image grayscale and grometric deformation in the Mixup approch. We demonstrate its effectiveness in medical image segmentation. This approach improves the accuracy of medical image segmentation.
Authors
- Peizhi Chen
- Gan Gao (ORCID: https://orcid.org/0000-0002-8100-7735)
Institutions
- Xiamen University of Technology (CN)
Publication Details
- Published
- 2024-09-13
- DOI
- https://doi.org/10.1109/itme63426.2024.00159
- Primary Topic
- Medical Image Segmentation Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00