Skip to main content
. 2020 May 6;10(5):806. doi: 10.3390/ani10050806

Figure 3.

Figure 3

(a) Anchor boxes generated around each of nine anchors [15]. (b) Inception V2 module in which the 5 × 5 convolutions are replaced by two successive 3 × 3 convolutions and pooling is applied [20].