def get_ssd(phase, num_classes):
vgg, extra_layers = add_vgg(3), add_extras(1024)
loc_layers = []
conf_layers = []
vgg_source = [21, -2]
for k, v in enumerate(vgg_source):
loc_layers += [
nn.Conv2d(vgg[v].out_channels,
mbox[k] * 4,
kernel_size=3,
padding=1)
]
conf_layers += [
nn.Conv2d(vgg[v].out_channels,
mbox[k] * num_classes,
kernel_size=3,
padding=1)
]
for k, v in enumerate(extra_layers[1::2], 2):
loc_layers += [
nn.Conv2d(v.out_channels, mbox[k] * 4, kernel_size=3, padding=1)
]
conf_layers += [
nn.Conv2d(v.out_channels,
mbox[k] * num_classes,
kernel_size=3,
padding=1)
]
SSD_MODEL = SSD(phase, vgg, extra_layers, (loc_layers, conf_layers),
num_classes)
return SSD_MODEL
def get_ssd(phase, num_classes):
vgg, extra_layers = add_vgg(3), add_extras(1024)
loc_layers = []
conf_layers = []
vgg_source = [21, -2]
for k, v in enumerate(
vgg_source): # k是索引，v是枚举出来得对象(0 21) (1 -2)提取出21层和33层对应（4，6）
loc_layers += [
nn.Conv2d(vgg[v].out_channels,
mbox[k] * 4,
kernel_size=3,
padding=1)
]
conf_layers += [
nn.Conv2d(vgg[v].out_channels,
mbox[k] * num_classes,
kernel_size=3,
padding=1)
]
for k, v in enumerate(extra_layers[1::2], 2): # （2，1）（3，3）（4，5）（5，7）
loc_layers += [
nn.Conv2d(v.out_channels, mbox[k] * 4, kernel_size=3, padding=1)
]
conf_layers += [
nn.Conv2d(v.out_channels,
mbox[k] * num_classes,
kernel_size=3,
padding=1)
]
SSD_MODEL = SSD(phase, vgg, extra_layers, (loc_layers, conf_layers),
num_classes)
return SSD_MODEL
def get_ssd(phase, num_classes, confidence=0.5, nms_iou=0.45):
#---------------------------------------------------#
# add_vgg指的是加入vgg主干特征提取网络。
# 该网络的最后一个特征层是conv7后的结果。
# shape为19,19,1024。
#
# 为了更好的提取出特征用于预测。
# SSD网络会继续进行下采样。
# add_extras是额外下采样的部分。
#---------------------------------------------------#
vgg, extra_layers = add_vgg(3), add_extras(1024)
loc_layers = []
conf_layers = []
vgg_source = [21, -2]
#---------------------------------------------------#
# 在add_vgg获得的特征层里
# 第21层和-2层可以用来进行回归预测和分类预测。
# 分别是conv4-3(38,38,512)和conv7(19,19,1024)的输出
#---------------------------------------------------#
for k, v in enumerate(vgg_source):
loc_layers += [
nn.Conv2d(vgg[v].out_channels,
mbox[k] * 4,
kernel_size=3,
padding=1)
]
conf_layers += [
nn.Conv2d(vgg[v].out_channels,
mbox[k] * num_classes,
kernel_size=3,
padding=1)
]
#-------------------------------------------------------------#
# 在add_extras获得的特征层里
# 第1层、第3层、第5层、第7层可以用来进行回归预测和分类预测。
# shape分别为(10,10,512), (5,5,256), (3,3,256), (1,1,256)
#-------------------------------------------------------------#
for k, v in enumerate(extra_layers[1::2], 2):
loc_layers += [
nn.Conv2d(v.out_channels, mbox[k] * 4, kernel_size=3, padding=1)
]
conf_layers += [
nn.Conv2d(v.out_channels,
mbox[k] * num_classes,
kernel_size=3,
padding=1)
]
#-------------------------------------------------------------#
# add_vgg和add_extras，一共获得了6个有效特征层，shape分别为：
# (38,38,512), (19,19,1024), (10,10,512),
# (5,5,256), (3,3,256), (1,1,256)
#-------------------------------------------------------------#
SSD_MODEL = SSD(phase, vgg, extra_layers, (loc_layers, conf_layers),
num_classes, confidence, nms_iou)
return SSD_MODEL
def __init__(self, num_classes, backbone_name, pretrained = False):
super(SSD300, self).__init__()
self.num_classes = num_classes
if backbone_name == "vgg":
self.vgg = add_vgg(pretrained)
self.extras = add_extras(1024, backbone_name)
self.L2Norm = L2Norm(512, 20)
mbox = [4, 6, 6, 6, 4, 4]
loc_layers = []
conf_layers = []
backbone_source = [21, -2]
#---------------------------------------------------#
# 在add_vgg获得的特征层里
# 第21层和-2层可以用来进行回归预测和分类预测。
# 分别是conv4-3(38,38,512)和conv7(19,19,1024)的输出
#---------------------------------------------------#
for k, v in enumerate(backbone_source):
loc_layers += [nn.Conv2d(self.vgg[v].out_channels, mbox[k] * 4, kernel_size = 3, padding = 1)]
conf_layers += [nn.Conv2d(self.vgg[v].out_channels, mbox[k] * num_classes, kernel_size = 3, padding = 1)]
#-------------------------------------------------------------#
# 在add_extras获得的特征层里
# 第1层、第3层、第5层、第7层可以用来进行回归预测和分类预测。
# shape分别为(10,10,512), (5,5,256), (3,3,256), (1,1,256)
#-------------------------------------------------------------#
for k, v in enumerate(self.extras[1::2], 2):
loc_layers += [nn.Conv2d(v.out_channels, mbox[k] * 4, kernel_size = 3, padding = 1)]
conf_layers += [nn.Conv2d(v.out_channels, mbox[k] * num_classes, kernel_size = 3, padding = 1)]
else:
self.mobilenet = mobilenet_v2(pretrained).features
self.extras = add_extras(1280, backbone_name)
self.L2Norm = L2Norm(96, 20)
mbox = [6, 6, 6, 6, 6, 6]
loc_layers = []
conf_layers = []
backbone_source = [13, -1]
for k, v in enumerate(backbone_source):
loc_layers += [nn.Conv2d(self.mobilenet[v].out_channels, mbox[k] * 4, kernel_size = 3, padding = 1)]
conf_layers += [nn.Conv2d(self.mobilenet[v].out_channels, mbox[k] * num_classes, kernel_size = 3, padding = 1)]
for k, v in enumerate(self.extras, 2):
loc_layers += [nn.Conv2d(v.out_channels, mbox[k] * 4, kernel_size = 3, padding = 1)]
conf_layers += [nn.Conv2d(v.out_channels, mbox[k] * num_classes, kernel_size = 3, padding = 1)]
self.loc = nn.ModuleList(loc_layers)
self.conf = nn.ModuleList(conf_layers)
self.backbone_name = backbone_name
def get_ssd(phase,num_classes):
#获得这些层
vgg, extra_layers = add_vgg(3), add_extras(1024)#in chnnels
loc_layers = []
conf_layers = []
vgg_source = [21, -2]#取出conv4-3进行回归预测，分类预测
for k, v in enumerate(vgg_source):# 0,21,,,,,,1,-2
loc_layers += [nn.Conv2d(vgg[v].out_channels,
mbox[k] * 4, kernel_size=3, padding=1)]#*4是因为中心点坐标2个，长宽2个，共可能调整的参数有4个,回归预测
conf_layers += [nn.Conv2d(vgg[v].out_channels,
mbox[k] * num_classes, kernel_size=3, padding=1)]# * num_classes是分类的种类，mbox[k]是在小格内预先设定k种不同尺寸anchor,分类预测
for k, v in enumerate(extra_layers[1::2], 2):
loc_layers += [nn.Conv2d(v.out_channels, mbox[k]
* 4, kernel_size=3, padding=1)]# 回归预测
conf_layers += [nn.Conv2d(v.out_channels, mbox[k]
* num_classes, kernel_size=3, padding=1)]# 分类预测
SSD_MODEL = SSD(phase, vgg, extra_layers, (loc_layers, conf_layers), num_classes)
return SSD_MODEL