[Retracted] Efficient Object Detection Algorithm in Kitchen Appliance Scene Images Based on Deep Learning

Figure 7

Structural diagram of four variants we set: (a) module only using 3 × 3 convolution layer; (b) module only using 5 × 5 convolution layer; (c) module only using 7 × 7 convolution layer; (d) our proposed multiscale module which adds 1 × 1 convolution layers following 3 × 3 convolution layer, 5 × 5 convolution layer, and 7 × 7 convolution layer.

(a)

(b)

(c)

(d)