Abstract
A serious issue with deep learning in medical applications is the limited availability of labeled medical image data. One of the data argumentation approach, Generative Adversarial Networks (GANs), can generate new samples that are drawn from the distribution of original dataset. Unfortunately, small sample medical imaging data is always not sufficient to train GANs with millions of parameters. Therefore, this paper proposes a pre-trained Style-based Generative Adversarial Networks (StyleGAN) to transfer knowledge from the Magnetic Resonance Imaging (MRI) domain to Computed tomography (CT) domain with limited sample images. This first pre-trained StyleGAN based Data augmentations (DA) method can generate high quality \(256\,\times \,256\) CT artifacts and artifact-free images for CT motion artifacts detection. Furthermore, we demonstrate this technique on a CT motion artifacts classification task and have achieved an improvement of \(5.59\%\) in sensitivity using synthetic images. Our results indicate that pre-trained models can provide priori knowledge to overcome the small sample problem in medical image processing.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Anwar, S.M., Majid, M., Qayyum, A., Awais, M., Alnowami, M., Khan, M.K.: Medical image analysis using convolutional neural networks: a review. J. Med. Syst. 42(11), 1–13 (2018). https://doi.org/10.1007/s10916-018-1088-1
Tajbakhsh, N., et al.: Convolutional neural networks for medical image analysis: full training or fine tuning? IEEE Trans. Med. Imaging 35(5), 1299–1312 (2016)
Wang, Y., Wu, C., Herranz, L., van de Weijer, J., Gonzalez-Garcia, A., Raducanu, B.: Transferring GANs: generating images from limited data. In: Proceedings of the European Conference on Computer Vision (ECCV), pp. 218–234 (2018)
Boas, F.E., Fleischmann, D.: CT artifacts: causes and reduction techniques. Imaging Med. 4(2), 229–240 (2012)
Goodfellow, I.J., et al.: Generative adversarial nets. Adv. Neural Inf. Process. Syst. 27, 2672–2680 (2014)
Karras, T., Laine, S., Aila, T.: A style-based generator architecture for generative adversarial networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4401-4410 (2019)
Chuquicusma, M. J., Hussein, S., Burt, J., Bagci, U.: How to fool radiologists with generative adversarial networks? A visual turing test for lung cancer diagnosis. In: 2018 IEEE 15th International Symposium on Biomedical Imaging, pp. 240–244 (2018)
Han, C., Murao, K., Noguchi, T., Kawata, Y., Uchiyama, F., Rundo, L., Satoh, S.I.: Learning more with less: conditional PGGAN-based data augmentation for brain metastases detection using highly-rough annotation on MR images. In: Proceedings of the 28th ACM International Conference on Information and Knowledge Management, pp. 119–127 (2019)
Ben-Cohen, A., et al.: Cross-modality synthesis from CT to PET using FCN and GAN networks for improved automated lesion detection. Eng. Appl. Artif. Intell. 4, 186–194 (2019)
Li, X., Grandvalet, Y., Davoine, F. Explicit inductive bias for transfer learning with convolutional networks. arXiv preprint arXiv:1802.01483 (2018)
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Su, K., Zhou, E., Sun, X., Wang, C., Yu, D., Luo, X. (2020). Pre-trained StyleGAN Based Data Augmentation for Small Sample Brain CT Motion Artifacts Detection. In: Yang, X., Wang, CD., Islam, M.S., Zhang, Z. (eds) Advanced Data Mining and Applications. ADMA 2020. Lecture Notes in Computer Science(), vol 12447. Springer, Cham. https://doi.org/10.1007/978-3-030-65390-3_26
Download citation
DOI: https://doi.org/10.1007/978-3-030-65390-3_26
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-65389-7
Online ISBN: 978-3-030-65390-3
eBook Packages: Computer ScienceComputer Science (R0)