Highly Robust Synthetic Aperture Radar Target Recognition Method Based on Simulation Data Training

Table 2

Architectures of VGGNets. The added layers are shown in bold, and the ReLU is placed after each weight layer, which is not shown for brevity.

VGG-11	VGG-13	VGG-16	VGG-16	VGG-19
Input image
Conv@3–64	Conv@3–64 Conv@3–64	Conv@3–64 Conv@3–64	Conv@3–64 Conv@3–64	Conv@3–64 Conv@3–64
Max pool, 2 × 2, s2
Conv@3–128	Conv@3–128 Conv@3–128	Conv@3–128 Conv@3–128	Conv@3–128 Conv@3–128	Conv@3–128 Conv@3–128
Max pool, 2 × 2, s2
Conv@3–256 Conv@3–256	Conv@3–256 Conv@3–256	Conv@3–256 Conv@3–256 Conv@1–256	Conv@3–256 Conv@3–256 Conv@3–256	Conv@3–256 Conv@3–256 Conv@3–256 Conv@3–256
Max pool, 2 × 2, s2
Conv@3–512 Conv@3–512	Conv@3–512 Conv@3–512	Conv@3–512 Conv@3–512 Conv@1–512	Conv@3–512 Conv@3–512 Conv@3–512	Conv@3–512 Conv@3–512 Conv@3–512 Conv@3–512
Max pool, 2 × 2, s2
Conv@3–512 Conv@3–512	Conv@3–512 Conv@3–512	Conv@3–512 Conv@3–512 Conv@1–512	Conv@3–512 Conv@3–512 Conv@3–512	Conv@3–512 Conv@3–512 Conv@3–512 Conv@3–512
Max pool, 2 × 2, s2
FC, 4096
FC, 4096
FC, 3
Softmax