Vision-Based Multiscale Construction Object Detection under Limited Supervision

Figure 4

Receptive field comparison of (a) conventional convolution and (b) DCN.

(a)

(b)