Open Access

Fig. 7

image

Download original image

Illustration of our architecture. This U-Net-like architecture comprises 11 PCDown Blocks in the encoder segment and an additional 11 PCUp Blocks in the decoder segment. The corresponding feature vectors at layer i in the encoder are copied and concatenated to the input of layer n - i in the decoder, where n is the total number of layers. Further details of the architecture and its parameters can be found in Table 4.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.