Skip to main content
Fig. 3 | Human-centric Computing and Information Sciences

Fig. 3

From: An improved object detection algorithm based on multi-scaled and deformable convolutional neural networks

Fig. 3

a Using an image pyramid to build a feature pyramid. b Recent detection systems have opted to use only single scale features for faster detection. c An alternative is to reuse the pyramidal feature hierarchy computed by ConvNet, as if it were a featurized image pyramid. d The feature pyramid network (FPN) is fast, as in b and c, but more accurate due to the fusion and multiple detection under different scaled features. The feature maps are indicated by blue outlines and thicker outlines denote semantically stronger features [20]

Back to article page