Echocardiography Segmentation by Quality Translation Using Anatomically Constrained CycleGAN
Segmentation of an echocardiogram (echo) is favorable for assessment of cardiac functionality and disease. The quality of the captured echo is a key factor that affects the segmentation accuracy. In this paper, we propose a novel generative adversarial network architecture, which aims to improve echo quality for the segmentation of the left ventricle (LV). The proposed model is anatomically constrained to the structure of the LV in apical four chamber (AP4) echo view. A set of discriminative features are learned through unpaired translation of low to high quality echo using adversarial training. The anatomical constraint regularizes the model during end-to-end training to preserve the corresponding shape of the LV in the translated echo. Experiments show that leveraging information in the translated high quality echocardiograms by the proposed method improves the robustness of the segmentation, where the worst-case Dice similarity score is improved by a margin of 15% over the baseline.
KeywordsAdversarial networks Image translation Quality improvement Segmentation Echocardiography
This work is supported in part by the Canadian Institutes of Health Research (CIHR) and in part by the Natural Sciences and Engineering Research Council of Canada (NSERC). The authors would like to acknowledge the support provided by Dale Hawley and the Vancouver Coastal Health in providing us with the anonymized, deidentified data.
- 1.Abdi, A.H., et al.: Automatic quality assessment of echocardiograms using convolutional neural networks: feasibility on the apical four-chamber view. IEEE TMI 36(6), 1221–1230 (2017)Google Scholar
- 2.Degel, M.A., Navab, N., Albarqouni, S.: Domain and geometry agnostic CNNs for left atrium segmentation in 3D ultrasound. In: Frangi, A.F., Schnabel, J.A., Davatzikos, C., Alberola-López, C., Fichtinger, G. (eds.) MICCAI 2018. LNCS, vol. 11073, pp. 630–637. Springer, Cham (2018). https://doi.org/10.1007/978-3-030-00937-3_72CrossRefGoogle Scholar
- 3.Huo, Y., et al.: Adversarial synthesis learning enables segmentation without target modality ground truth. In: IEEE ISBI, pp. 1217–1220 (2018)Google Scholar
- 4.Leclerc, S., et al.: Deep learning for segmentation using an open large-scale dataset in 2D echocardiography. IEEE TMI 38, 2198–2210 (2019)Google Scholar
- 5.Noble, J.A., Boukerroui, D.: Ultrasound image segmentation: a survey. IEEE TMI 25(8), 987–1010 (2006)Google Scholar
- 6.Oktay, O., et al.: Anatomically constrained neural networks (ACNNs): application to cardiac image enhancement and segmentation. IEEE TMI 37(2), 384–395 (2018)Google Scholar
- 7.Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-24574-4_28CrossRefGoogle Scholar
- 8.Salimans, T., et al.: Improved techniques for training GANs. In: NIPS, pp. 2234–2242 (2016)Google Scholar
- 10.Zhang, Z., et al.: Translating and segmenting multimodal medical volumes with cycle- and shape-consistency generative adversarial network. In: IEEE CVPR (2018)Google Scholar
- 11.Zhu, J.Y., et al.: Unpaired image-to-image translation using cycle-consistent adversarial networks. In: IEEE CVPR, pp. 2223–2232 (2017)Google Scholar