def __init__(self,
down_conv_nn,
use_xyz,
bn=True,
activation=nn.ReLU(),
*args,
**kwargs):
super(RSConvMapper, self).__init__()
self._down_conv_nn = down_conv_nn
self._use_xyz = use_xyz
self.nn = nn.ModuleDict()
if len(self._down_conv_nn) == 2: # First layer
self._first_layer = True
f_in, f_intermediate, f_out = self._down_conv_nn[0]
self.nn["features_nn"] = pt_utils.SharedMLP(self._down_conv_nn[1],
bn=bn)
else:
self._first_layer = False
f_in, f_intermediate, f_out = self._down_conv_nn
self.nn["mlp_msg"] = pt_utils.SharedMLP([f_in, f_intermediate, f_out],
bn=bn)
self.nn["norm"] = Seq(*[nn.BatchNorm2d(f_out), activation])
self._f_out = f_out
def __init__(self,
down_conv_nn,
channel_raising_nn,
use_xyz,
bn=True,
*args,
**kwargs):
super(RSConvMapper, self).__init__()
self._down_conv_nn = down_conv_nn
self._channel_raising_nn = channel_raising_nn
self._use_xyz = use_xyz
if len(self._down_conv_nn) == 2: # First layer
self._first_layer = True
f_in, f_intermediate, f_out = self._down_conv_nn[0]
self.features_nn = pt_utils.SharedMLP(self._down_conv_nn[1], bn=bn)
else:
self._first_layer = False
f_in, f_intermediate, f_out = self._down_conv_nn
self.mlp_msg = pt_utils.SharedMLP([f_in, f_intermediate, f_out], bn=bn)
self.mlp_out = nn.Conv1d(f_out,
channel_raising_nn[-1],
kernel_size=(1, 1))
def __init__(self,
npoint=None,
radii=None,
nsample=None,
down_conv_nn=None,
bn=True,
activation="LeakyReLU",
use_xyz=True,
**kwargs):
assert len(radii) == len(nsample) == len(down_conv_nn)
super(PointNetMSGDown,
self).__init__(DenseFPSSampler(num_to_sample=npoint),
DenseRadiusNeighbourFinder(radii, nsample),
**kwargs)
self.use_xyz = use_xyz
self.npoint = npoint
self.mlps = nn.ModuleList()
for i in range(len(radii)):
mlp_spec = down_conv_nn[i]
if self.use_xyz:
mlp_spec[0] += 3
self.mlps.append(
pt_utils.SharedMLP(down_conv_nn[i],
bn=bn,
activation=get_activation(activation)))
def __init__(self,
*,
npoint: int,
radii: List[float],
nsamples: List[int],
mlps: List[List[int]],
bn: bool = True,
use_xyz: bool = True):
super().__init__()
assert len(radii) == len(nsamples) == len(mlps)
self.npoint = npoint
self.groupers = nn.ModuleList()
self.mlps = nn.ModuleList()
for i in range(len(radii)):
radius = radii[i]
nsample = nsamples[i]
self.groupers.append(
pointnet2_utils.QueryAndGroup(radius, nsample, use_xyz=use_xyz)
if npoint is not None else pointnet2_utils.GroupAll(use_xyz))
mlp_spec = mlps[i]
if use_xyz:
mlp_spec[0] += 3
self.mlps.append(pt_utils.SharedMLP(mlp_spec, bn=bn))
def __init__(self,
radii,
nsamples,
mlps,
bn=True,
use_xyz=True,
vote=False):
# type: (PointnetSAModuleMSG, int, list[float], list[int], list[list[int]], bool, bool) -> None
super(PointnetSAModuleMSG, self).__init__()
# 继承了_PointnetSAModuleBase中定义的操作方法
assert len(radii) == len(nsamples) == len(mlps)
self.groupers = nn.ModuleList()
self.mlps = nn.ModuleList()
for i in range(len(radii)):
radius = radii[i] # 多尺度就在于球查询有多种半径
nsample = nsamples[i]
if vote is False:
self.groupers.append(
pointnet2_utils.QueryAndGroup(radius,
nsample,
use_xyz=use_xyz))
# self.add_module(str(len(self)), module)
else: # 如果要霍夫投票，就要计算得分
self.groupers.append(
pointnet2_utils.QueryAndGroup_score(radius,
nsample,
use_xyz=use_xyz))
mlp_spec = mlps[i]
if use_xyz: # 如果要使用坐标信息，则MLP的输入要加三个
mlp_spec[0] += 3
self.mlps.append(pt_utils.SharedMLP(mlp_spec, bn=bn))
def __init__(self,
radii,
nsamples,
mlps,
bn=True,
use_xyz=True,
vote=False):
# type: (PointnetSAModuleMSG, int, List[float], List[int], List[List[int]], bool, bool) -> None
super(PointnetSAModuleMSG, self).__init__()
assert len(radii) == len(nsamples) == len(mlps)
self.groupers = nn.ModuleList()
self.mlps = nn.ModuleList()
for i in range(len(radii)):
radius = radii[i]
nsample = nsamples[i]
if vote is False:
self.groupers.append(
pointnet2_utils.QueryAndGroup(radius,
nsample,
use_xyz=use_xyz))
else:
self.groupers.append(
pointnet2_utils.QueryAndGroup_score(radius,
nsample,
use_xyz=use_xyz))
mlp_spec = mlps[i]
if use_xyz:
mlp_spec[0] += 3
self.mlps.append(pt_utils.SharedMLP(mlp_spec, bn=bn))
def __init__(self,
up_conv_nn,
bn=True,
bias=False,
activation="LeakyReLU",
**kwargs):
super(DenseFPModule, self).__init__(None, **kwargs)
self.nn = pt_utils.SharedMLP(up_conv_nn,
bn=bn,
activation=get_activation(activation))
def __init__(self,
nn,
aggr="max",
bn=True,
activation="LeakyReLU",
**kwargs):
super(GlobalDenseBaseModule, self).__init__()
self.nn = pt_utils.SharedMLP(nn,
bn=bn,
activation=get_activation(activation))
if aggr.lower() not in ["mean", "max"]:
raise Exception(
"The aggregation provided is unrecognized {}".format(aggr))
self._aggr = aggr.lower()
def __init__(self, npoint, radii, nsamples, mlps, bn=True):
# type: (PointnetSAModuleMSGRRI, int, List[float], List[int], List[List[int]], bool, bool) -> None
super(PointnetSAModuleMSGRRI, self).__init__()
assert len(radii) == len(nsamples) == len(mlps)
self.npoint = npoint
self.groupers = nn.ModuleList()
self.mlps = nn.ModuleList()
for i in range(len(radii)):
radius = radii[i]
nsample = nsamples[i]
self.groupers.append(
pointnet2_utils.QueryAndGroupRRI(radius, nsample))
mlp_spec = mlps[i]
self.mlps.append(pt_utils.SharedMLP(mlp_spec, bn=bn))
def __init__(self, input_channels=3, use_xyz=True, objective=False):
super(Pointnet_Tracking, self).__init__()
self.backbone_net = Pointnet_Backbone(input_channels, use_xyz)
self.cosine = nn.CosineSimilarity(dim=1)
self.mlp = pt_utils.SharedMLP([4 + 256, 256, 256, 256], bn=True)
self.FC_layer_cla = (pt_utils.Seq(256).conv1d(256, bn=True).conv1d(
256, bn=True).conv1d(1, activation=None))
self.fea_layer = (pt_utils.Seq(256).conv1d(256, bn=True).conv1d(
256, activation=None))
self.vote_layer = (pt_utils.Seq(3 + 256).conv1d(256, bn=True).conv1d(
256, bn=True).conv1d(3 + 256, activation=None))
self.vote_aggregation = PointnetSAModule(radius=0.3,
nsample=16,
mlp=[1 + 256, 256, 256, 256],
use_xyz=use_xyz)
self.num_proposal = 64
self.FC_proposal = (pt_utils.Seq(256).conv1d(256, bn=True).conv1d(
256, bn=True).conv1d(3 + 1 + 1, activation=None))
def __init__(self, mlp, bn=True):
# type: (PointnetFPModule, List[int], bool) -> None
super(PointnetFPModule, self).__init__()
self.mlp = pt_utils.SharedMLP(mlp, bn=bn)
def __init__(self, input_channels=3, use_xyz=True, objective=False):
super(Pointnet_Tracking, self).__init__()
self.backbone_net = Pointnet_Backbone(input_channels, use_xyz)
self.cosine = nn.CosineSimilarity(dim=1)
self.mlp = pt_utils.SharedMLP([4 + 256, 256, 256, 256], bn=True)
# SharedMLP(
# (layer0): Conv2d(
# (conv): Conv2d(260, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
# (normlayer): BatchNorm2d(
# (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (layer1): Conv2d(
# (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
# (normlayer): BatchNorm2d(
# (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (layer2): Conv2d(
# (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
# (normlayer): BatchNorm2d(
# (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# )
self.FC_layer_cla = (pt_utils.Seq(256).conv1d(256, bn=True).conv1d(
256, bn=True).conv1d(1, activation=None))
# Seq(
# (0): Conv1d(
# (conv): Conv1d(256, 256, kernel_size=(1,), stride=(1,), bias=False)
# (normlayer): BatchNorm1d(
# (bn): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (1): Conv1d(
# (conv): Conv1d(256, 256, kernel_size=(1,), stride=(1,), bias=False)
# (normlayer): BatchNorm1d(
# (bn): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (2): Conv1d(
# (conv): Conv1d(256, 1, kernel_size=(1,), stride=(1,))
# )
# )
self.fea_layer = (pt_utils.Seq(256).conv1d(256, bn=True).conv1d(
256, activation=None))
# Seq(
# (0): Conv1d(
# (conv): Conv1d(256, 256, kernel_size=(1,), stride=(1,), bias=False)
# (normlayer): BatchNorm1d(
# (bn): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (1): Conv1d(
# (conv): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
# )
# )
self.vote_layer = (pt_utils.Seq(3 + 256).conv1d(256, bn=True).conv1d(
256, bn=True).conv1d(3 + 256, activation=None))
# Seq(
# (0): Conv1d(
# (conv): Conv1d(259, 256, kernel_size=(1,), stride=(1,), bias=False)
# (normlayer): BatchNorm1d(
# (bn): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (1): Conv1d(
# (conv): Conv1d(256, 256, kernel_size=(1,), stride=(1,), bias=False)
# (normlayer): BatchNorm1d(
# (bn): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (2): Conv1d(
# (conv): Conv1d(256, 259, kernel_size=(1,), stride=(1,))
# )
# )
self.vote_aggregation = PointnetSAModule(
radius=0.3,
nsample=16,
mlp=[1 + 256, 256, 256, 256],
use_xyz=use_xyz) # 根据投票聚合投影候选中心点
# PointnetSAModule(
# (groupers): ModuleList(
# (0): QueryAndGroup()
# )
# (mlps): ModuleList(
# (0): SharedMLP(
# (layer0): Conv2d(
# (conv): Conv2d(260, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
# (normlayer): BatchNorm2d(
# (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (layer1): Conv2d(
# (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
# (normlayer): BatchNorm2d(
# (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# (layer2): Conv2d(
# (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
# (normlayer): BatchNorm2d(
# (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (activation): ReLU(inplace=True)
# )
# )
# )
# )
self.num_proposal = 64 # 实际上是_PointnetSAModuleBase(nn.Module)的forward中的npoint输入参数
self.FC_proposal = (pt_utils.Seq(256).conv1d(256, bn=True).conv1d(
256, bn=True).conv1d(3 + 1 + 1, activation=None))
def __init__(self, nn, **kwargs):
super(GlobalDenseBaseModule, self).__init__()
self.nn = pt_utils.SharedMLP(nn)
def __init__(self, up_conv_nn, bn=True, **kwargs):
super(DenseFPModule, self).__init__(None, **kwargs)
self.nn = pt_utils.SharedMLP(up_conv_nn, bn=bn)
def __init__(self, *, mlp: List[int], bn: bool = True):
super().__init__()
self.mlp = pt_utils.SharedMLP(mlp, bn=bn)
