目标检测 SSD: Single Shot MultiBox Detector - VGG16的魔改

目标检测 SSD: Single Shot MultiBox Detector - VGG16的魔改

flyish

看图

前部分是VGG16，VGG16中的FC6和FC7改成了卷积层，后部分是SSD自定义的层，自定义的层叫Extra Feature Layders

到detection的箭头一共有6个

相当于从6层卷积层中get features.
特征大小是38*38、19*19、10*10、5*5、3*3、1*1
以下6层分别是

38×38 
Conv4_3

19×19
Conv7

10×10
Conv8_2

5×5 
Conv9_2

3×3 
Conv10_2

1×1 
Conv11_2

VGG16中的片段

(26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(27): ReLU(inplace)
(28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(29): ReLU(inplace)
(30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)

SSD根据上面的30往下接是(相当于FC6和FC7改成了卷积层)

(31): Conv2d(512, 1024, kernel_size=(3, 3), stride=(1, 1), padding=(6, 6), dilation=(6, 6))
(32): ReLU(inplace)
(33): Conv2d(1024, 1024, kernel_size=(1, 1), stride=(1, 1))
(34): ReLU(inplace)

VGG16去掉了

  (classifier): Sequential(
    (0): Linear(in_features=25088, out_features=4096, bias=True)
    (1): ReLU(inplace)
    (2): Dropout(p=0.5)
    (3): Linear(in_features=4096, out_features=4096, bias=True)
    (4): ReLU(inplace)
    (5): Dropout(p=0.5)
    (6): Linear(in_features=4096, out_features=1000, bias=True)
  )

default box
特征大小3838、1919、1010、55、33、11
自定规则的 4, 6, 6, 6, 4, 4

（38*38*4 + 19*19*6 + 10*10*6 + 5*5*6 + 3*3*4 + 1*1*4）= 8732个default box


  (extras): ModuleList(
    (0): Sequential(
      (0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
      (3): ReLU()
    )
    (1): Sequential(
      (0): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
      (3): ReLU()
    )
    (2): Sequential(
      (0): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1))
      (3): ReLU()
    )
    (3): Sequential(
      (0): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1))
      (3): ReLU()
    )
  )

Detection之前完整的就是

SSD(
  (base_net): ModuleList(
    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU(inplace)
    (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU(inplace)
    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (6): ReLU(inplace)
    (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (8): ReLU(inplace)
    (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (11): ReLU(inplace)
    (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (13): ReLU(inplace)
    (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (15): ReLU(inplace)
    (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=True)
    (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (18): ReLU(inplace)
    (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (20): ReLU(inplace)
    (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (22): ReLU(inplace)
    (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (25): ReLU(inplace)
    (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (27): ReLU(inplace)
    (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (29): ReLU(inplace)
    (30): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)
    (31): Conv2d(512, 1024, kernel_size=(3, 3), stride=(1, 1), padding=(6, 6), dilation=(6, 6))
    (32): ReLU(inplace)
    (33): Conv2d(1024, 1024, kernel_size=(1, 1), stride=(1, 1))
    (34): ReLU(inplace)
  )
  (extras): ModuleList(
    (0): Sequential(
      (0): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
      (3): ReLU()
    )
    (1): Sequential(
      (0): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
      (3): ReLU()
    )
    (2): Sequential(
      (0): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1))
      (3): ReLU()
    )
    (3): Sequential(
      (0): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1))
      (1): ReLU()
      (2): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1))
      (3): ReLU()
    )
  )
）

VGG16原始的

(30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)

SSD改装VGG16的

(30): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=False)

分类 20+1=21 （ 4, 6, 6, 6, 4, 4）

(0): Conv2d(512, 84, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): Conv2d(1024, 126, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(2): Conv2d(512, 126, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(3): Conv2d(256, 126, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): Conv2d(256, 84, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(5): Conv2d(256, 84, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))

位置（ 4, 6, 6, 6, 4, 4）每个都乘以4（cx,cy,w,h）

(0): Conv2d(512, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): Conv2d(1024, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(2): Conv2d(512, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(3): Conv2d(256, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): Conv2d(256, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(5): Conv2d(256, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))