Skip to main content
EJNMMI Physics
Fig. 1 | EJNMMI Physics

Fig. 1

From: Automated quantification of baseline imaging PET metrics on FDG PET/CT images of pediatric Hodgkin lymphoma patients

Fig. 1

DeepMedic network design, adapted for lymph node detection. The number of features in all convolutional layers was 90, 90, 110, 110, 110, 110, 130, and 130. The number of features in the two fully-connected layers was 250. The input segment size of the downsampled pathways are set so that the feature maps prior to the final connected layer are of equivalent size, as in [22].

Back to article page