Skip to main content
. 2020 May 20;10(5):330. doi: 10.3390/diagnostics10050330

Figure 2.

Figure 2

The RetinaNet architecture uses a Feature Pyramid Network [21] backbone on top of a feedforward ResNet architecture [22] (a) to generate a rich, multi-scale convolutional feature pyramid (b). To this backbone RetinaNet attaches two subnetworks: one for classifying anchor boxes (c) and one for regressing from anchor boxes to ground-truth object boxes (d). (Reprinted and adapted with permission from ICCV 2017 [20].)