Skip to main content
. Author manuscript; available in PMC: 2021 Jun 8.
Published in final edited form as: Domain Adapt Represent Transf Med Image Learn Less Labels Imperfect Data (2019). 2019 Oct 13;2019:54–62. doi: 10.1007/978-3-030-33391-1_7

Fig. 2.

Fig. 2

Qualitative results on a single slice from a single subject in the POPPY dataset. The top row shows a slice from the in-plane FLAIR acquisition whilst the bottom row shows a slice from the volumetric FLAIR acquisition. Each column shows a model’s predictions on that row’s image. This slice is used to highlight an example of an artefact (shown in the red circle) introduced by the in-plane acquisition. The baseline method introduces a false positive in this region whilst the domain adaptation methods perform better at ignoring it. Our approach shows the best in-plane to volumetric agreement.