Skip to main content
. Author manuscript; available in PMC: 2019 Sep 1.
Published in final edited form as: Med Image Comput Comput Assist Interv. 2018 Sep 26;11071:921–929. doi: 10.1007/978-3-030-00934-2_102

Fig. 3.

Fig. 3.

Two generic FCN architectures: DN and UN with three layer aggregation structures: (a) deep supervision (DS), (b) deeply supervised feature pyramid network (DS-FPN), and (c) deeply supervised iterative deep aggregation (DS-IDA). Intuitively, the box size is proportional to the spatial resolution and the box linewidth to the feature channel number.