Figure 3.

The structure of our proposed backbone network PFH-PSA. The red arrows indicate deconvolution layers, purple boxes indicate $1\times 1$ convolutional layers, light yellow boxes indicate $3\times 3$ convolutional layers, orange boxes indicate $5\times 5$ convolutional layers, brown boxes indicate $7\times 7$ convolutional layers, and blue boxes indicate the convolutional layers with stride = 2. Different colors indicate feature maps with different scales, and different textures indicate feature maps obtained through deconvolution layers or max-pooling layers.