def get_edgeconv(point_cloud,
k,
mlp,
is_training,
bn_decay,
scope,
bn=True,
associated=None,
is_dist=False):
with tf.variable_scope(scope) as sc:
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
if associated is not None:
for j, feature in enumerate(associated):
point_cloud = tf.concat([point_cloud, feature], axis=-1)
edge_feature = tf_util.get_edge_feature(point_cloud,
nn_idx=nn_idx,
k=k)
for i, num_out_channel in enumerate(mlp):
edge_feature = tf_util.conv2d(edge_feature,
num_out_channel, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='conv%d' % (i),
bn_decay=bn_decay,
is_dist=is_dist)
edge_feature = tf.reduce_max(edge_feature, axis=-2, keep_dims=True)
return edge_feature
def get_sampled_edgeconv(point_cloud,
sampled_point_cloud,
k,
mlp,
is_training,
bn_decay,
scope,
bn=True,
is_dist=False):
with tf.variable_scope(scope) as sc:
adj_matrix = tf_util.sampled_pairwise_distance(sampled_point_cloud,
point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_sampled_edge_feature(sampled_point_cloud,
point_cloud,
nn_idx=nn_idx,
k=k)
for i, num_out_channel in enumerate(mlp):
edge_feature = tf_util.conv2d(edge_feature,
num_out_channel, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='conv%d' % (i),
bn_decay=bn_decay,
is_dist=is_dist)
edge_feature = tf.reduce_max(edge_feature, axis=-2, keep_dims=True)
return edge_feature
def get_model_old_4(point_cloud, is_training, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
batch_size = point_cloud.get_shape()[0].value
# point_cloud_2_momentum = tf.concat(
# [tf.multiply(point_cloud, point_cloud), tf.multiply(tf.roll(point_cloud, shift=1, axis=-1), point_cloud)],
# axis=-1)
num_point = point_cloud.get_shape()[1].value
end_points = {}
k = 20
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.new_get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 128, [1, 1],
padding='VALID', stride=[1, 1],
bn=True, is_training=is_training,
scope='dgcnn1', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
# MLP on global point cloud vector
net = tf.reshape(net, [batch_size, -1])
net = tf_util.fully_connected(net, 512, bn=True, is_training=is_training,
scope='fc1', bn_decay=bn_decay)
net = tf_util.dropout(net, keep_prob=0.5, is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net, end_points
def knn_loss(self, input_point_cloud, K=10):
adj_matrix = tf_util.pairwise_distance(input_point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=K)
knn_distances = tf_util.get_neighbor_distances(input_point_cloud,
nn_idx=nn_idx,
k=K)
knn_loss_val = tf.reduce_mean(knn_distances, [1, 2])
return knn_loss_val
def get_sampled_edgeconv_groupconv(point_cloud,
sampled_point_cloud,
k,
mlp,
is_training,
bn_decay,
scope,
bn=True,
sampled_pc=None,
pc=None,
is_dist=False):
with tf.variable_scope(scope) as sc:
adj_matrix = tf_util.sampled_pairwise_distance(sampled_point_cloud,
point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
if pc is not None:
point_cloud = tf.concat(
[point_cloud, tf.expand_dims(pc, axis=-2)], axis=-1)
if sampled_pc is not None:
sampled_point_cloud = tf.concat(
[sampled_point_cloud,
tf.expand_dims(sampled_pc, axis=-2)],
axis=-1)
edge_feature, neighbors, concat = tf_util.get_sampled_edge_feature_separate(
sampled_point_cloud, point_cloud, nn_idx=nn_idx, k=k)
for i, num_out_channel in enumerate(mlp):
edge_feature = tf_util.conv2d(edge_feature,
num_out_channel // 2, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='edge_conv%d' % (i),
bn_decay=bn_decay,
is_dist=is_dist)
neighbors = tf_util.conv2d(neighbors,
num_out_channel // 2, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='neig_conv%d' % (i),
bn_decay=bn_decay,
is_dist=is_dist)
net = channle_shuffle(edge_feature, neighbors)
if i < len(mlp) - 1:
ch = edge_feature.get_shape().as_list()[-1]
edge_feature = net[:, :, :, 0:ch]
neighbors = net[:, :, :, ch:]
net = tf.reduce_max(net, axis=-2, keep_dims=True)
return net
def get_edgeconv_groupconv(point_cloud,
k,
mlp,
is_training,
bn_decay,
scope,
bn=True,
associated=None):
with tf.variable_scope(scope) as sc:
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
if associated is not None:
for j, feature in enumerate(associated):
point_cloud = tf.concat([point_cloud, feature], axis=-1)
net = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
for i, num_out_channel in enumerate(mlp):
center, edge_feature = tf.split(net, num_or_size_splits=2, axis=-1)
center = tf_util.conv2d(center,
num_out_channel, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='centerconv%d' % (i),
bn_decay=bn_decay)
edge_feature = tf_util.conv2d(edge_feature,
num_out_channel, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='edgeconv%d' % (i),
bn_decay=bn_decay)
net = channle_shuffle(center, edge_feature)
edge_feature = tf.reduce_max(net, axis=-2, keep_dims=True)
return edge_feature
def same_label_topology(point_cloud, seg_label, k):
# point_cloud bn3
# seg_label bn50
# adj = tf_util.pairwise_distance(point_cloud)
# nn_idx = tf_util.knn(adj, k=k)
#print('point_cloud',point_cloud.shape)
#print('seg_label',seg_label)
seg_label = tf.to_float(seg_label)
adj_seg = tf_util.pairwise_distance(seg_label) #bnn
adj_pc = tf_util.pairwise_distance(point_cloud) #bnn
max_dist = tf.reduce_max(tf.reduce_max(adj_pc, axis=-1), axis=-1)
adj_sl = adj_pc + adj_seg * max_dist
#print('adj_pc',adj_pc.shape,adj_pc)
#print('adj_sl',adj_sl.shape)
nn_idx = tf_util.knn(adj_sl, k=k)
return nn_idx
def get_model(points,
n_iter,
is_training,
bn_decay,
randinit=False,
nostop=False,
k=10):
T = tf.eye(4, batch_shape=(points.shape[0], ))
T_deltas = []
for i in range(n_iter):
transformed_points = tf_util.transform_points(points, T)
if not nostop:
transformed_points = tf.stop_gradient(transformed_points)
adj_matrix = tf_util.pairwise_distance(transformed_points)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(points, nn_idx=nn_idx, k=k)
qt = input_transform_net(edge_feature, is_training, bn_decay, randinit)
T_delta = tf.map_fn(tf_util.qt2mat, qt, dtype=tf.float32)
T_deltas.append(T_delta)
T = tf.matmul(T_delta, T)
transformed_points = tf_util.transform_points(points, T)
return transformed_points, T, T_deltas
def model_part(point_cloud, is_training, k, bn_decay=None):
out1 = tf_util.conv2d(point_cloud, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='samp_conv1', bn_decay=bn_decay, is_dist=True)
out2 = tf_util.conv2d(out1, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='samp_conv2', bn_decay=bn_decay, is_dist=True)
net_max_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
net_mean_1 = tf.reduce_mean(out2, axis=-2, keep_dims=True)
out3 = tf_util.conv2d(tf.concat([net_max_1, net_mean_1], axis=-1), 64, [1,1],
padding='VALID', stride=[2,1],
bn=True, is_training=is_training,
scope='samp_conv3', bn_decay=bn_decay, is_dist=True)
print("out3 = ", out3.shape)
out3_max = tf.reduce_max(out3, axis=1, keep_dims=True)
print("out3_max = ", out3_max.shape)
adj = tf_util.pairwise_distance(tf.squeeze(out3, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(out3, nn_idx=nn_idx, k=k)
out4 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='samp_conv4', bn_decay=bn_decay, is_dist=True)
net_max_2 = tf.reduce_max(out4, axis=-2, keep_dims=True)
net_mean_2 = tf.reduce_mean(out4, axis=-2, keep_dims=True)
out5 = tf_util.conv2d(tf.concat([net_max_2, net_mean_2], axis=-1), 64, [1,1],
padding='VALID', stride=[2,1],
bn=True, is_training=is_training,
scope='samp_conv5', bn_decay=bn_decay, is_dist=True)
out5_max = tf.reduce_max(out5, axis=1, keep_dims=True)
adj = tf_util.pairwise_distance(tf.squeeze(out5, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(out5, nn_idx=nn_idx, k=k)
out6 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='samp_conv6', bn_decay=bn_decay, is_dist=True)
net_max_3 = tf.reduce_max(out6, axis=-2, keep_dims=True)
net_mean_3 = tf.reduce_mean(out6, axis=-2, keep_dims=True)
out7 = tf_util.conv2d(tf.concat([net_max_3, net_mean_3], axis=-1), 64, [1,1],
padding='VALID', stride=[2,1],
bn=True, is_training=is_training,
scope='samp_conv7', bn_decay=bn_decay, is_dist=True)
out7_max = tf.reduce_max(out7, axis=1, keep_dims=True)
out8 = tf_util.conv2d(out7, 1024, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='samp_conv8', bn_decay=bn_decay, is_dist=True)
print("out8 = ", out8.shape)
out_max = tf.reduce_max(out8, axis=1, keep_dims=True)
print("out_max = ", out_max.shape)
return out3_max, out5_max, out7_max, out_max
def get_model(point_cloud,
is_training,
num_class,
global_pl,
params,
weight_decay=None,
bn_decay=None,
scname=''):
''' input: BxNxF
Use https://arxiv.org/pdf/1902.08570 as baseline
output:BxNx(cats*segms) '''
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -2)
k = params[0]
#adj = tf_util.pairwise_distance(point_cloud[:,:,:3])
adj = tf_util.pairwise_distanceR(point_cloud[:, :, :3])
n_heads = params[1]
nn_idx = tf_util.knn(adj, k=k)
net, locals_transform, coefs = gap_block(k, n_heads, nn_idx, point_cloud,
point_cloud,
('filter0', params[2]), bn_decay,
weight_decay, is_training, scname)
net = tf_util.conv2d(net,
params[3], [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='gapnet00',
bn_decay=bn_decay)
net00 = net
net = tf_util.conv2d(net,
params[4], [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet01' + scname,
bn_decay=bn_decay)
net01 = net
net = tf_util.conv2d(net,
params[5], [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet02' + scname,
bn_decay=bn_decay)
net02 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
adj_conv = nn_idx
n_heads = params[6]
net, locals_transform1, coefs2 = gap_block(k, n_heads, nn_idx, net,
point_cloud,
('filter1', params[7]),
bn_decay, weight_decay,
is_training, scname)
net = tf_util.conv2d(net,
params[8], [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='gapnet10',
bn_decay=bn_decay)
net10 = net
net = tf_util.conv2d(net,
params[9], [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet11' + scname,
bn_decay=bn_decay)
net11 = net
net = tf_util.conv2d(net,
params[10], [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet12' + scname,
bn_decay=bn_decay)
net12 = net
global_expand = tf.reshape(global_pl, [batch_size, 1, 1, -1])
global_expand = tf.tile(global_expand, [1, num_point, 1, 1])
global_expand = tf_util.conv2d(global_expand,
16, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='global_expand' + scname,
bn_decay=bn_decay)
net = tf.concat([
net00, net01, net02, net11, net12, global_expand, locals_transform,
locals_transform1
],
axis=-1)
net = tf_util.conv2d(net,
params[8], [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='agg' + scname,
bn_decay=bn_decay)
#net_tot = net
net = tf_util.avg_pool2d(net, [num_point, 1],
padding='VALID',
scope='avgpool' + scname)
expand = tf.tile(net, [1, num_point, 1, 1])
# net = tf.concat(axis=3, values=[expand,
# net_tot,
# ])
net = tf_util.conv2d(expand,
params[11], [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv2',
weight_decay=weight_decay,
is_dist=True)
net = tf_util.dropout(net,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp1')
net = tf_util.conv2d(net,
params[11], [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv3',
weight_decay=weight_decay,
is_dist=True)
net = tf_util.dropout(net,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp2')
net = tf_util.conv2d(net,
params[12], [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv4',
weight_decay=weight_decay,
is_dist=True)
net = tf_util.conv2d(net,
num_class, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=None,
bn=False,
scope='seg/conv5',
weight_decay=weight_decay,
is_dist=True)
net = tf.cond(is_training, lambda: net, lambda: tf.nn.softmax(net))
net = tf.reshape(net, [batch_size, num_point, num_class])
return net, coefs, coefs2, adj_conv
def calc_ldgcnn_feature(point_cloud, is_training, bn_decay=None):
# B: batch size; N: number of points, C: channels; k: number of nearest neighbors
# point_cloud: B*N*3
k = 20
# adj_matrix: B*N*N
adj_matrix = tf_util.pairwise_distance(point_cloud)
# Find the indices of knearest neighbors.
nn_idx = tf_util.knn(adj_matrix, k=k)
point_cloud = tf.expand_dims(point_cloud, axis=-2)
# Edge_feature: B*N*k*6
# The vector in the last dimension represents: (Xc,Yc,Zc, Xck - Xc, Yck-Yc, Yck-zc)
# (Xc,Yc,Zc) is the central point. (Xck - Xc, Yck-Yc, Yck-zc) is the edge vector.
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
# net: B*N*k*64
# The kernel size of CNN is 1*1, and thus this is a MLP with sharing parameters.
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn1',
bn_decay=bn_decay)
# net: B*N*1*64
# Extract the biggest feature from k convolutional edge features.
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net1 = net
# adj_matrix: B*N*N
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
# net: B*N*1*67
# Link the Hierarchical features.
net = tf.concat([point_cloud, net1], axis=-1)
# edge_feature: B*N*k*134
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
# net: B*N*k*64
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn2',
bn_decay=bn_decay)
# net: B*N*1*64
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net2 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
# net: B*N*1*131
net = tf.concat([point_cloud, net1, net2], axis=-1)
# edge_feature: B*N*k*262
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
# net: B*N*k*64
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn3',
bn_decay=bn_decay)
# net: B*N*1*64
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net3 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
# net: B*N*1*195
net = tf.concat([point_cloud, net1, net2, net3], axis=-1)
# edge_feature: B*N*k*390
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
# net: B*N*k*128
net = tf_util.conv2d(edge_feature,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn4',
bn_decay=bn_decay)
# net: B*N*1*128
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net4 = net
# input: B*N*1*323
# net: B*N*1*1024
net = tf_util.conv2d(tf.concat([point_cloud, net1, net2, net3, net4],
axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay)
# net: B*1*1*1024
net = tf.reduce_max(net, axis=1, keep_dims=True)
# net: B*1024
net = tf.squeeze(net)
return net
def get_model(point_cloud, is_training, bn_decay=None):
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
end_points = {}
k = 20
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
n_heads = 1
attns = []
local_features = []
for i in range(n_heads):
edge_feature, coefs, locals = attn_feature(point_cloud,
16,
nn_idx,
activation=tf.nn.elu,
in_dropout=0.6,
coef_dropout=0.6,
is_training=is_training,
bn_decay=bn_decay,
layer='layer0',
k=k,
i=i)
attns.append(edge_feature)
local_features.append(locals)
neighbors_features = tf.concat(attns, axis=-1)
neighbors_features = tf.concat(
[tf.expand_dims(point_cloud, -2), neighbors_features], axis=-1)
locals_max_transform = tf.reduce_max(tf.concat(local_features, axis=-1),
axis=-2,
keep_dims=True)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(neighbors_features,
locals_max_transform,
is_training,
bn_decay,
K=3)
point_cloud_transformed = tf.matmul(point_cloud, transform)
adj_matrix = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj_matrix, k=k)
n_heads = 4
attns = []
local_features = []
for i in range(n_heads):
edge_feature, coefs, locals = attn_feature(point_cloud_transformed,
16,
nn_idx,
activation=tf.nn.elu,
in_dropout=0.6,
coef_dropout=0.6,
is_training=is_training,
bn_decay=bn_decay,
layer='layer1',
k=k,
i=i)
attns.append(edge_feature)
local_features.append(locals)
neighbors_features = tf.concat(attns, axis=-1)
neighbors_features = tf.concat(
[tf.expand_dims(point_cloud_transformed, -2), neighbors_features],
axis=-1)
net = tf_util.conv2d(neighbors_features,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='gapnet1',
bn_decay=bn_decay)
net1 = net
locals_max = tf.reduce_max(tf.concat(local_features, axis=-1),
axis=-2,
keep_dims=True)
net = tf_util.conv2d(net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='gapnet2',
bn_decay=bn_decay)
net2 = net
net = tf_util.conv2d(net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='gapnet3',
bn_decay=bn_decay)
net3 = net
net = tf_util.conv2d(net,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='gapnet4',
bn_decay=bn_decay)
net4 = net
net = tf_util.conv2d(tf.concat([net1, net2, net3, net4, locals_max],
axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=1, keep_dims=True)
net = tf.reshape(net, [batch_size, -1])
net = tf_util.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net, end_points
def get_model(point_cloud, input_label, is_training, cat_num, part_num, \
batch_size, num_point, weight_decay=.00004, bn_decay=None):
bn_decay = bn_decay if bn_decay is not None else 0.9
with tf.variable_scope("DGCNN"):
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)
k = 20
bn_params = {
"is_training": is_training,
"decay": bn_decay,
'renorm': True
}
adj = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image,
nn_idx=nn_idx,
k=k)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3)
point_cloud_transformed = tf.matmul(point_cloud, transform)
input_image = tf.expand_dims(point_cloud_transformed, -1)
adj = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image,
nn_idx=nn_idx,
k=k)
# out1 = tf_util.conv2d(edge_feature, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv1', bn_decay=bn_decay, is_dist=True)
out1 = layers.masked_conv2d(
edge_feature,
64,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='adj_conv1')
# out2 = tf_util.conv2d(out1, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv2', bn_decay=bn_decay, is_dist=True)
out2 = layers.masked_conv2d(
out1,
64,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='adj_conv2')
net_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
adj = tf_util.pairwise_distance(net_1)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_1, nn_idx=nn_idx, k=k)
# out3 = tf_util.conv2d(edge_feature, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv3', bn_decay=bn_decay, is_dist=True)
out3 = layers.masked_conv2d(
edge_feature,
64,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='adj_conv3')
# out4 = tf_util.conv2d(out3, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv4', bn_decay=bn_decay, is_dist=True)
out4 = layers.masked_conv2d(
out3,
64,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='adj_conv4')
net_2 = tf.reduce_max(out4, axis=-2, keep_dims=True)
adj = tf_util.pairwise_distance(net_2)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_2, nn_idx=nn_idx, k=k)
# out5 = tf_util.conv2d(edge_feature, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv5', bn_decay=bn_decay, is_dist=True)
out5 = layers.masked_conv2d(
edge_feature,
64,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='adj_conv5')
# out6 = tf_util.conv2d(out5, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv6', bn_decay=bn_decay, is_dist=True)
net_3 = tf.reduce_max(out5, axis=-2, keep_dims=True)
# out7 = tf_util.conv2d(tf.concat([net_1, net_2, net_3], axis=-1), 1024, [1, 1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training,
# scope='adj_conv7', bn_decay=bn_decay, is_dist=True)
out7 = layers.masked_conv2d(
tf.concat([net_1, net_2, net_3], axis=-1),
1024,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='adj_conv7')
# out_max = tf_util.max_pool2d(out7, [num_point, 1], padding='VALID', scope='maxpool')
out_max = slim.max_pool2d(out7, [num_point, 1],
stride=1,
padding='VALID',
scope='maxpool')
one_hot_label_expand = tf.reshape(input_label,
[batch_size, 1, 1, cat_num])
# one_hot_label_expand = tf_util.conv2d(one_hot_label_expand, 64, [1, 1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training,
# scope='one_hot_label_expand', bn_decay=bn_decay, is_dist=True)
one_hot_label_expand = layers.masked_conv2d(
one_hot_label_expand,
64,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='one_hot_label_expand')
out_max = tf.concat(axis=3, values=[out_max, one_hot_label_expand])
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3, values=[expand, net_1, net_2, net_3])
# net2 = tf_util.conv2d(concat, 256, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
# bn=True, is_training=is_training, scope='seg/conv1', weight_decay=weight_decay, is_dist=True)
net2 = layers.masked_conv2d(
concat,
256,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='seg/conv1')
# net2 = tf_util.dropout(net2, keep_prob=0.6, is_training=is_training, scope='seg/dp1')
net2 = slim.dropout(net2,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp1')
# net2 = tf_util.conv2d(net2, 256, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
# bn=True, is_training=is_training, scope='seg/conv2', weight_decay=weight_decay, is_dist=True)
net2 = layers.masked_conv2d(
net2,
256,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='seg/conv2')
# net2 = tf_util.dropout(net2, keep_prob=0.6, is_training=is_training, scope='seg/dp2')
net2 = slim.dropout(net2,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp2')
# net2 = tf_util.conv2d(net2, 128, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
# bn=True, is_training=is_training, scope='seg/conv3', weight_decay=weight_decay, is_dist=True)
net2 = layers.masked_conv2d(
net2,
128,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=slim.batch_norm,
normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=tf.nn.relu6,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='seg/conv3')
# net2 = tf_util.conv2d(net2, part_num, [1,1], padding='VALID', stride=[1,1], activation_fn=None,
# bn=False, scope='seg/conv4', weight_decay=weight_decay, is_dist=True)
net2 = layers.masked_conv2d(
net2,
part_num,
# max(int(round(64 * scale)), 32),
[1, 1],
padding='VALID',
stride=1,
normalizer_fn=None,
# normalizer_params=bn_params,
biases_initializer=tf.contrib.layers.xavier_initializer(),
weights_regularizer=slim.l2_regularizer(weight_decay),
activation_fn=None,
weights_initializer=tf.contrib.layers.xavier_initializer(),
scope='seg/conv4')
net2 = tf.reshape(net2, [batch_size, num_point, part_num])
return net2
def get_model(point_cloud, input_label, is_training, cat_num, part_num, \
batch_size, num_point, weight_decay, graphnum, featnum, bn_decay=None):
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)
k = 30
adj = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3,
is_dist=True)
point_cloud_transformed = tf.matmul(point_cloud, transform)
input_image = tf.expand_dims(point_cloud_transformed, -2)
adj = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
out1 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv1',
bn_decay=bn_decay,
is_dist=True)
out2 = tf_util.conv2d(out1,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv2',
bn_decay=bn_decay,
is_dist=True)
net_max_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
net_mean_1 = tf.reduce_mean(out2, axis=-2, keep_dims=True)
out3 = tf_util.conv2d(tf.concat([net_max_1, net_mean_1], axis=-1),
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv3',
bn_decay=bn_decay,
is_dist=True)
out3A, net_max_1A, net_mean_1A = tf_util.offset_deform(
input_image,
out3,
scope="trans_conv0",
num_neighbor=k,
num_graph=graphnum[0],
num_feat=featnum[0],
weight_decay=weight_decay,
is_training=is_training,
bn_decay=bn_decay)
out5, net_max_2, net_mean_2 = tf_util.offset_deform(
input_image,
out3A,
scope="trans_conv1",
num_neighbor=k,
num_graph=graphnum[0],
num_feat=featnum[0],
weight_decay=weight_decay,
is_training=is_training,
bn_decay=bn_decay)
out7, net_max_3, net_mean_3 = tf_util.offset_deform(
input_image,
out5,
scope="trans_conv2",
num_neighbor=k,
num_graph=graphnum[1],
num_feat=featnum[1],
weight_decay=weight_decay,
is_training=is_training,
bn_decay=bn_decay)
'''adj = tf_util.pairwise_distance(tf.squeeze(trans2, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(tf.concat([out5,trans2], axis = -1), nn_idx=nn_idx, k=k)
out6 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv6', bn_decay=bn_decay, is_dist=True)
net_max_3 = tf.reduce_max(out6, axis=-2, keep_dims=True)
net_mean_3 = tf.reduce_mean(out6, axis=-2, keep_dims=True)
out7 = tf_util.conv2d(tf.concat([net_max_3, net_mean_3], axis=-1), 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv7', bn_decay=bn_decay, is_dist=True)'''
out8 = tf_util.conv2d(tf.concat([out3, out5, out7], axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='adj_conv13',
bn_decay=bn_decay,
is_dist=True)
out_max = tf_util.max_pool2d(out8, [num_point, 1],
padding='VALID',
scope='maxpool')
one_hot_label_expand = tf.reshape(input_label, [batch_size, 1, 1, cat_num])
one_hot_label_expand = tf_util.conv2d(one_hot_label_expand,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='one_hot_label_expand',
bn_decay=bn_decay,
is_dist=True)
out_max = tf.concat(axis=3, values=[out_max, one_hot_label_expand])
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3,
values=[
expand, net_max_1, net_mean_1, out3, net_max_2,
net_mean_2, out5, net_max_3, net_mean_3, out7, out8
])
net2 = tf_util.conv2d(concat,
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv1',
weight_decay=weight_decay,
is_dist=True)
net2 = tf_util.dropout(net2,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp1')
net2 = tf_util.conv2d(net2,
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv2',
weight_decay=weight_decay,
is_dist=True)
net2 = tf_util.dropout(net2,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp2')
net2 = tf_util.conv2d(net2,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv3',
weight_decay=weight_decay,
is_dist=True)
net2 = tf_util.conv2d(net2,
part_num, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=None,
bn=False,
scope='seg/conv4',
weight_decay=weight_decay,
is_dist=True)
net2 = tf.reshape(net2, [batch_size, num_point, part_num])
return net2
def get_model(point_cloud,
is_training,
num_class,
weight_decay=None,
bn_decay=None,
scname=''):
''' input: BxNxF
output:BxNx(cats*segms) '''
batch_size = point_cloud.get_shape()[0]
num_point = point_cloud.get_shape()[1]
num_feat = point_cloud.get_shape()[2]
k = 10
adj = tf_util.pairwise_distanceR(point_cloud[:, :, :3])
n_heads = 1
nn_idx = tf_util.knn(adj, k=k)
net, locals_transform, coefs = gap_block(k, n_heads, nn_idx, point_cloud,
point_cloud, ('filter0', 16),
bn_decay, weight_decay,
is_training, scname)
net = tf_util.conv2d(net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet01' + scname,
bn_decay=bn_decay)
net01 = net
net = tf_util.conv2d(net,
128, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet02' + scname,
bn_decay=bn_decay)
net02 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
adj_conv = nn_idx
n_heads = 1
net, locals_transform1, coefs2 = gap_block(k, n_heads, nn_idx, net,
point_cloud, ('filter1', 128),
bn_decay, weight_decay,
is_training, scname)
net = tf_util.conv2d(net,
256, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet11' + scname,
bn_decay=bn_decay)
net11 = net
net = tf_util.conv2d(net,
256, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet12' + scname,
bn_decay=bn_decay)
net12 = net
net = tf.concat(
[net01, net02, net11, net12, locals_transform, locals_transform1],
axis=-1)
net = tf_util.conv2d(net,
3, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='agg' + scname,
bn_decay=bn_decay)
net = tf_util.avg_pool2d(net, [num_point, 1],
padding='VALID',
scope='avgpool' + scname)
max_pool = net
net = tf.reshape(net, [batch_size, -1])
net = tf_util.fully_connected(net,
256,
bn=True,
is_training=is_training,
activation_fn=tf.nn.relu,
scope='fc1' + scname,
bn_decay=bn_decay)
net = tf_util.fully_connected(net,
128,
bn=True,
is_training=is_training,
activation_fn=tf.nn.relu,
scope='fc2' + scname,
bn_decay=bn_decay)
net = tf_util.fully_connected(net,
num_class,
activation_fn=None,
scope='fc3' + scname)
net = tf.squeeze(net)
return net, max_pool
def get_model(point_cloud, is_training, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
end_points = {}
k = 20
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
print(adj_matrix, nn_idx, edge_feature)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature, is_training, bn_decay, K=3)
point_cloud_transformed = tf.matmul(point_cloud, transform)
adj_matrix = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud_transformed, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn1', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net1 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn2', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net2 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn3', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net3 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 128, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn4', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net4 = net
net = tf_util.conv2d(tf.concat([net1, net2, net3, net4], axis=-1), 1024, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='agg', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=1, keep_dims=True)
# MLP on global point cloud vector
end_points['post_max'] = net
net = tf.reshape(net, [batch_size, -1])
net = tf_util.fully_connected(net, 512, bn=True, is_training=is_training,
scope='fc1', bn_decay=bn_decay)
net = tf_util.dropout(net, keep_prob=0.5, is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net, 256, bn=True, is_training=is_training,
scope='fc2', bn_decay=bn_decay)
net = tf_util.dropout(net, keep_prob=0.5, is_training=is_training,
scope='dp2')
end_points['final'] = net
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net, end_points
def get_model_w_ae_gcn(self, point_cloud, is_training, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
tf_util = imp.load_source(
'tf_util',
os.path.join(os.path.dirname(self.models["test"]), '../utils',
"tf_util.py"))
transform_nets = imp.load_source(
'transform_nets',
os.path.join(os.path.dirname(self.models["test"]),
"transform_nets.py"))
import tf_util
from transform_nets import input_transform_net
batch_size = self.configuration.batch_size
num_point = self.configuration.n_input[0]
end_points = {}
k = 20
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud,
nn_idx=nn_idx,
k=k)
print(adj_matrix, nn_idx, edge_feature)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3)
point_cloud_transformed = tf.matmul(point_cloud, transform)
adj_matrix = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud_transformed,
nn_idx=nn_idx,
k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn1',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net1 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn2',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net2 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn3',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net3 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn4',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net4 = net
net = tf_util.conv2d(tf.concat([net1, net2, net3, net4], axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=1, keep_dims=True)
# MLP on global point cloud vector
net = tf.reshape(net, [batch_size, -1])
end_points['post_max'] = net
net = tf_util.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
end_points['final'] = net
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net, end_points
def forward(point_cloud, is_training, bn_decay=None):
"""LPD-Net:FNSF, INPUT is batch_num_queries X num_pointclouds_per_query X num_points_per_pointcloud X 13,
OUTPUT batch_num_queries X num_pointclouds_per_query X output_dim """
batch_num_queries = point_cloud.get_shape()[0].value
num_pointclouds_per_query = point_cloud.get_shape()[1].value
num_points = point_cloud.get_shape()[2].value
CLUSTER_SIZE=64
OUTPUT_DIM=256
k=20
point_cloud = tf.reshape(point_cloud, [batch_num_queries*num_pointclouds_per_query, num_points,13])
point_cloud, feature_cloud = tf.split(point_cloud, [3,10], 2)
with tf.variable_scope('transform_net1') as sc:
input_transform = input_transform_net(point_cloud, is_training, bn_decay, K=3)
point_cloud_transformed = tf.matmul(point_cloud, input_transform)
# Neural Network to learn neighborhood features
# feature_cloud = neural_feature_net(point_cloud, is_training, bn_decay, knn_k=20, F=10)
point_cloud_input = tf.concat([point_cloud_transformed, feature_cloud], 2)
point_cloud_input = tf.expand_dims(point_cloud_input, -1)
net = tf_util.conv2d(point_cloud_input, 64, [1, 13],
padding='VALID', stride=[1,1],
is_training=is_training,
scope='conv1', bn_decay=bn_decay)
net = tf_util.conv2d(net, 64, [1, 1],
padding='VALID', stride=[1,1],
is_training=is_training,
scope='conv2', bn_decay=bn_decay)
with tf.variable_scope('transform_net2') as sc:
feature_transform = feature_transform_net(net, is_training, bn_decay, K=64)
feature_transform = tf.matmul(tf.squeeze(net, axis=[2]), feature_transform)
# Serial structure
# Danymic Graph cnn for feature space
with tf.variable_scope('DGfeature') as sc:
adj_matrix = tf_util.pairwise_distance(feature_transform)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1, 1],
padding='VALID', stride=[1, 1],
bn=True, is_training=is_training,
scope='dgmlp1', bn_decay=bn_decay)
net = tf_util.conv2d(net, 64, [1, 1],
padding='VALID', stride=[1, 1],
bn=True, is_training=is_training,
scope='dgmlp2', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
# Spatial Neighborhood fusion for cartesian space
with tf.variable_scope('SNfeature') as sc:
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
idx_ = tf.range(batch_num_queries*num_pointclouds_per_query) * num_points
idx_ = tf.reshape(idx_, [batch_num_queries*num_pointclouds_per_query, 1, 1])
feature_cloud = tf.reshape(net, [-1, 64])
edge_feature = tf.gather(feature_cloud, nn_idx+idx_)
net = tf_util.conv2d(edge_feature, 64, [1, 1],
padding='VALID', stride=[1, 1],
bn=True, is_training=is_training,
scope='snmlp1', bn_decay=bn_decay)
net = tf_util.conv2d(net, 64, [1, 1],
padding='VALID', stride=[1, 1],
bn=True, is_training=is_training,
scope='snmlp2', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
# MLP for fusion
net = tf_util.conv2d(net, 64, [1, 1],
padding='VALID', stride=[1,1],
is_training=is_training,
scope='conv3', bn_decay=bn_decay)
net = tf_util.conv2d(net, 128, [1, 1],
padding='VALID', stride=[1,1],
is_training=is_training,
scope='conv4', bn_decay=bn_decay)
net = tf_util.conv2d(net, 1024, [1, 1],
padding='VALID', stride=[1,1],
is_training=is_training,
scope='conv5', bn_decay=bn_decay)
point_wise_feature = net
NetVLAD = lp.NetVLAD(feature_size=1024, max_samples=num_points, cluster_size=CLUSTER_SIZE,
output_dim=OUTPUT_DIM, gating=True, add_batch_norm=True,
is_training=is_training)
net= tf.reshape(net,[-1,1024])
net = tf.nn.l2_normalize(net,1)
output = NetVLAD.forward(net)
print(output)
#normalize to have norm 1
output = tf.nn.l2_normalize(output,1)
output = tf.reshape(output,[batch_num_queries,num_pointclouds_per_query,OUTPUT_DIM])
return output
def model(point_cloud, is_training, cut, num_point, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
k = 20
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
cut,
bn_decay,
K=3)
point_cloud_transformed = tf.matmul(point_cloud, transform)
adj_matrix = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud_transformed,
nn_idx=nn_idx,
k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope=cut + 'dgcnn1',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net1 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope=cut + 'dgcnn2',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net2 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope=cut + 'dgcnn3',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net3 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope=cut + 'dgcnn4',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net4 = net
net = tf_util.conv2d(tf.concat([net1, net2, net3, net4], axis=-1),
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope=cut + 'agg',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=1, keep_dims=True)
return net
def get_model(point_cloud, is_training, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
end_points = {}
k = 20 # 近邻点
# 得到边的特征和点的特征
adj_matrix = tf_util.pairwise_distance(point_cloud) # 两两点之间的距离 B*N*N
nn_idx = tf_util.knn(adj_matrix, k=k) # B*N*20
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx,
k=k) # 输入为B*N*3,B*N*20，输出B*N*k*6
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3) # B*3*3
point_cloud_transformed = tf.matmul(point_cloud, transform) # 点云对齐
# dgcnn1
adj_matrix = tf_util.pairwise_distance(
point_cloud_transformed
) # 输入：batch_size, num_points, num_dims 输出：8*1024*1024
nn_idx = tf_util.knn(adj_matrix, k=k) # B*N*20
edge_feature = tf_util.get_edge_feature(point_cloud_transformed,
nn_idx=nn_idx,
k=k) # B*N*20*6=B*N*K*6
# 特征提取，卷积核为1*1，输出维度为64
net = tf_util.conv2d(
edge_feature,
64,
[1, 1], # B*N*K*64
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn1',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True) # 取最大 B*N*1*64
net1 = net
# dgcnn2
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx,
k=k) # B*N*K*128
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn2',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net2 = net
# dgcnn3
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn3',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net3 = net
# dgcnn4
adj_matrix = tf_util.pairwise_distance(net) # B*N*N
nn_idx = tf_util.knn(adj_matrix, k=k) # B*N*20
edge_feature = tf_util.get_edge_feature(
net, nn_idx=nn_idx, k=k) # 输入：net-B*N*1*64 8*1024*20*128
net = tf_util.conv2d(
edge_feature,
128,
[1, 1], # 8*1024*20*128
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn4',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True) # 8*1024*1*128
net4 = net
net = tf_util.conv2d(
tf.concat([net1, net2, net3, net4], axis=-1),
1024,
[1, 1], # 输出1024个特征，卷积核1*1
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay) # 8*1024*1*1024
net = tf.reduce_max(net, axis=1, keep_dims=True) # 8*1*1*128
# MLP on global point cloud vector
net = tf.reshape(net, [batch_size, -1]) # 8*1024
net = tf_util.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net, end_points
def get_model(point_cloud,
is_training,
num_class,
global_pl,
weight_decay=None,
bn_decay=None,
scname=''):
''' input: BxNxF
output:BxNx(cats*segms) '''
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
num_feat = point_cloud.get_shape()[2].value
k = 10
adj = tf_util.pairwise_distanceR(point_cloud[:, :, :3])
n_heads = 1
nn_idx = tf_util.knn(adj, k=k)
net, locals_transform, coefs = gap_block(k, n_heads, nn_idx, point_cloud,
point_cloud, ('filter0', 16),
bn_decay, weight_decay,
is_training, scname)
net = tf_util.conv2d(net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet01' + scname,
bn_decay=bn_decay)
net01 = net
net = tf_util.conv2d(net,
128, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet02' + scname,
bn_decay=bn_decay)
net02 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
adj_conv = nn_idx
n_heads = 1
net, locals_transform1, coefs2 = gap_block(k, n_heads, nn_idx, net,
point_cloud, ('filter1', 128),
bn_decay, weight_decay,
is_training, scname)
net = tf_util.conv2d(net,
256, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet11' + scname,
bn_decay=bn_decay)
net11 = net
net = tf_util.conv2d(net,
256, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='gapnet12' + scname,
bn_decay=bn_decay)
net12 = net
global_expand = tf.reshape(global_pl, [batch_size, 1, 1, -1])
global_expand = tf.tile(global_expand, [1, num_point, 1, 1])
global_expand = tf_util.conv2d(global_expand,
16, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='global_expand' + scname,
bn_decay=bn_decay)
net = tf.concat([
net01, net02, net11, net12, global_expand, locals_transform,
locals_transform1
],
axis=-1)
net = tf_util.conv2d(net,
2, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=tf.nn.relu,
bn=True,
is_training=is_training,
scope='agg' + scname,
bn_decay=bn_decay)
net = tf_util.max_pool2d(net, [num_point, 1],
padding='VALID',
scope='avgpool' + scname)
max_pool = net
expand = tf.tile(net, [1, num_point, 1, 1])
net = tf.concat(axis=3, values=[
expand,
net01,
net11,
])
net = tf_util.conv2d(net,
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv2',
weight_decay=weight_decay,
is_dist=True)
net = tf_util.conv2d(net,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv4',
weight_decay=weight_decay,
is_dist=True)
net = tf_util.conv2d(net,
num_class, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=None,
bn=False,
scope='seg/conv5',
weight_decay=weight_decay,
is_dist=True)
net = tf.reshape(net, [batch_size, num_point, num_class])
return net, max_pool
def get_model(point_cloud, is_training, bn_decay=None):
""" ConvNet baseline, input is BxNx9 gray image """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)
k = 30
adj = tf_util.pairwise_distance(point_cloud[:, :, 6:])
nn_idx = tf_util.knn(adj, k=k) # (batch, num_points, k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
samp_out1, samp_out2, samp_out3, globle_feat = model_part(edge_feature, is_training, k, bn_decay)
out1 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv1', bn_decay=bn_decay, is_dist=True)
out2 = tf_util.conv2d(out1, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv2', bn_decay=bn_decay, is_dist=True)
net_max_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
net_mean_1 = tf.reduce_mean(out2, axis=-2, keep_dims=True)
out3 = tf_util.conv2d(tf.concat([net_max_1, net_mean_1], axis=-1), 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv3', bn_decay=bn_decay, is_dist=True)
out1_expand = tf.tile(tf.reshape(samp_out1, [batch_size, 1, 1, -1]), [1, num_point, 1, 1])
out1_concat = tf.concat(axis=3, values=[out3, out1_expand])
print("out1_concat = ", out1_concat.shape)
out4 = tf_util.conv2d(out1_concat, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv4', bn_decay=bn_decay, is_dist=True)
adj = tf_util.pairwise_distance(tf.squeeze(out4, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(out4, nn_idx=nn_idx, k=k)
out5 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv5', bn_decay=bn_decay, is_dist=True)
net_max_2 = tf.reduce_max(out5, axis=-2, keep_dims=True)
net_mean_2 = tf.reduce_mean(out5, axis=-2, keep_dims=True)
out6 = tf_util.conv2d(tf.concat([net_max_2, net_mean_2], axis=-1), 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv6', bn_decay=bn_decay, is_dist=True)
out2_expand = tf.tile(tf.reshape(samp_out2, [batch_size, 1, 1, -1]), [1, num_point, 1, 1])
out2_concat = tf.concat(axis=3, values=[out6, out2_expand])
out7 = tf_util.conv2d(out2_concat, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv7', bn_decay=bn_decay, is_dist=True)
adj = tf_util.pairwise_distance(tf.squeeze(out7, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(out7, nn_idx=nn_idx, k=k)
out8 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv8', bn_decay=bn_decay, is_dist=True)
net_max_3 = tf.reduce_max(out8, axis=-2, keep_dims=True)
net_mean_3 = tf.reduce_mean(out8, axis=-2, keep_dims=True)
out9 = tf_util.conv2d(tf.concat([net_max_3, net_mean_3], axis=-1), 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv9', bn_decay=bn_decay, is_dist=True)
out3_expand = tf.tile(tf.reshape(samp_out3, [batch_size, 1, 1, -1]), [1, num_point, 1, 1])
out3_concat = tf.concat(axis=3, values=[out9, out3_expand])
out10 = tf_util.conv2d(out3_concat, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv10', bn_decay=bn_decay, is_dist=True)
out11 = tf_util.conv2d(tf.concat([out4, out7, out10], axis=-1), 1024, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv11', bn_decay=bn_decay, is_dist=True)
out_max = tf_util.max_pool2d(out11, [num_point,1], padding='VALID', scope='maxpool')
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3, values=[expand,
net_max_1,
net_mean_1,
out4,
net_max_2,
net_mean_2,
out7,
net_max_3,
net_mean_3,
out10,
out11])
print("concat = ", concat.shape)
# CONCAT
globle_feat_expand = tf.tile(tf.reshape(globle_feat, [batch_size, 1, 1, -1]), [1, num_point, 1, 1])
points_feat1_concat = tf.concat(axis=3, values=[concat, globle_feat_expand])
print("points_feat1_concat = ", points_feat1_concat.shape)
# CONV
net = tf_util.conv2d(points_feat1_concat, 512, [1,1], padding='VALID', stride=[1,1],
bn=True, is_training=is_training, scope='seg/conv1', is_dist=True)
net = tf_util.conv2d(net, 256, [1,1], padding='VALID', stride=[1,1],
bn=True, is_training=is_training, scope='seg/conv2', is_dist=True)
net = tf_util.dropout(net, keep_prob=0.7, is_training=is_training, scope='dp1')
net = tf_util.conv2d(net, 13, [1,1], padding='VALID', stride=[1,1],
activation_fn=None, scope='seg/conv3', is_dist=True)
net = tf.squeeze(net, [2])
return net
def get_model(input_tensor, is_training, bn_decay = None):
weight_decay = 0.0
num_point = input_tensor.get_shape()[1].value
k = 40
#Transform Net
adj_matrix = tf_util.pairwise_distance(input_tensor)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(input_tensor, nn_idx=nn_idx, k=k)
#Transform Net
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature, is_training, bn_decay, K=input_tensor.get_shape()[2], is_dist=True)
input_tensor_transformed = tf.matmul(input_tensor, transform)
adj_matrix = tf_util.pairwise_distance(input_tensor_transformed)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(input_tensor_transformed, nn_idx=nn_idx, k=k)
out1_1 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='one/adj_conv1', bn_decay=bn_decay, is_dist=True)
out1_2 = tf_util.conv2d(out1_1, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='one/adj_conv2', bn_decay=bn_decay, is_dist=True)
out1_3 = tf_util.conv2d(out1_2, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='one/adj_conv3', bn_decay=bn_decay, is_dist=True)
net_1 = tf.reduce_max(out1_3, axis=-2, keepdims=True)
adj = tf_util.pairwise_distance(net_1)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_1, nn_idx=nn_idx, k=k)
out2_1 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='two/adj_conv1', bn_decay=bn_decay, is_dist=True)
out2_2 = tf_util.conv2d(out2_1, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='two/adj_conv2', bn_decay=bn_decay, is_dist=True)
out2_3 = tf_util.conv2d(out2_2, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='two/adj_conv3', bn_decay=bn_decay, is_dist=True)
net_2 = tf.reduce_max(out2_3, axis=-2, keepdims=True)
adj = tf_util.pairwise_distance(net_2)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_2, nn_idx=nn_idx, k=k)
out3_1 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='three/adj_conv1', bn_decay=bn_decay, is_dist=True)
out3_2 = tf_util.conv2d(out3_1, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='three/adj_conv2', bn_decay=bn_decay, is_dist=True)
net_3 = tf.reduce_max(out3_2, axis=-2, keepdims=True)
out7 = tf_util.conv2d(tf.concat([net_1, net_2, net_3], axis=-1), 1024, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv7', bn_decay=bn_decay, is_dist=True)
out_max = tf_util.max_pool2d(out7, [num_point, 1], padding='VALID', scope='maxpool')
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3, values=[expand,
net_1,
net_2,
net_3])
# CONV
net = tf_util.conv2d(concat, 512, [1,1], padding='VALID', stride=[1,1],
bn=True, is_training=is_training, scope='seg/conv1', is_dist=True)
# net = tf_util.conv2d(net, 256, [1,1], padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, scope='seg/conv2', is_dist=True)
# net = tf_util.conv2d(net, 128, [1,1], padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, scope='seg/conv3', is_dist=True)
# net = tf_util.conv2d(net, 64, [1,1], padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, scope='seg/conv4', is_dist=True)
# net = tf_util.conv2d(net, 32, [1,1], padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, scope='seg/conv5', is_dist=True)
net = tf_util.dropout(net, keep_prob=0.5, is_training=is_training, scope='dp1')
net = tf_util.conv2d(net, 16, [1,1], padding='VALID', stride=[1,1],
activation_fn=None, scope='seg/output', is_dist=True)
net = tf.squeeze(net, [2])
return net
def get_model(point_cloud, is_training, bn_decay, k=20):
""" Classification DGCNN, input is BxNx3, output Bx40 """
# EdgeConv functions (MLP implemented as conv2d)
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn1', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keepdims=True)
net1 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn2', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keepdims=True)
net2 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn3', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keepdims=True)
net3 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature, 128, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='dgcnn4', bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keepdims=True)
net4 = net
net = tf_util.conv2d(tf.concat([net1, net2, net3, net4], axis=-1), 1024, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='agg', bn_decay=bn_decay)
net = tf.squeeze(net, -2)
# Symmetric function: max pooling
net = tf.reduce_max(net, axis=1, name='maxpool')
# MLP on global point cloud vector
net = tf_util.fully_connected(net, 512, bn=True, is_training=is_training,
scope='fc1', bn_decay=bn_decay)
net = tf_util.dropout(net, keep_prob=0.5, is_training=is_training, scope='dp1')
net = tf_util.fully_connected(net, 256, bn=True, is_training=is_training,
scope='fc2', bn_decay=bn_decay)
net = tf_util.dropout(net, keep_prob=0.5, is_training=is_training, scope='dp2')
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net
def get_model(point_cloud, is_training, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
end_points = {}
k = 20
adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3)
point_cloud_transformed = tf.matmul(point_cloud, transform)
input_image = tf.expand_dims(point_cloud_transformed, -1)
adj_matrix = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(point_cloud_transformed,
nn_idx=nn_idx,
k=k)
# addition of transform layers
net = tf_util.conv2d(input_image,
64, [1, 3],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='conv1',
bn_decay=bn_decay)
net = tf_util.conv2d(net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='conv2',
bn_decay=bn_decay)
with tf.variable_scope('transform_net2') as sc:
transform = feature_transform_net(net, is_training, bn_decay, K=64)
end_points['transform'] = transform
net_transformed = tf.matmul(tf.squeeze(net, axis=[2]), transform)
net_transformed = tf.expand_dims(net_transformed, [2])
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet1',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net1 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet2',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net2 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet3',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net3 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet4',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net4 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet5',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net5 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet6',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net6 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet6',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net7 = net
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dftnet7',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net8 = net
net = tf_util.conv2d(tf.concat(
[net1, net2, net3, net4, net5, net6, net7, net8], axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=1, keep_dims=True)
net = tf.reshape(net, [batch_size, -1])
net = tf_util.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
net = tf_util.fully_connected(net, 40, activation_fn=None, scope='fc3')
return net, end_points
def get_model(point_cloud, is_training, num_classes, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
batch_size = point_cloud.get_shape()[0].value
# num_point = point_cloud.get_shape()[1].value
end_points = {}
k = 20
# pairwise distance of the points in the point cloud
adj_matrix = tf_util.pairwise_distance(point_cloud)
# get indices of k nearest neighbors
nn_idx = tf_util.knn(adj_matrix, k=k)
# edge feature
edge_feature = tf_util.get_edge_feature(point_cloud, nn_idx=nn_idx, k=k)
# transform net 1
with tf.variable_scope('transform_net1') as _:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3)
# point cloud transf
point_cloud_transformed = tf.matmul(point_cloud, transform)
# pairwise distance of the points in the point cloud
adj_matrix = tf_util.pairwise_distance(point_cloud_transformed)
# get indices of k nearest neighbors
nn_idx = tf_util.knn(adj_matrix, k=k)
# I've got neighbors indices and subregion index (0-7)
edge_feature = tf_util.get_edge_feature(point_cloud_transformed,
nn_idx=nn_idx,
k=k)
# net = tf_util.conv2d_reg(point_cloud_transformed, nn_idx,
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn1',
bn_decay=bn_decay)
# Maxpool
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net_1 = net
#############################################################################
# 2nd block
#############################################################################
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn2',
bn_decay=bn_decay)
# net = tf.reduce_max(net, axis=-2, keep_dims=False)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net_2 = net
#############################################################################
# 3rd block
#############################################################################
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn3',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net_3 = net
#############################################################################
# 4rd block
#############################################################################
adj_matrix = tf_util.pairwise_distance(net)
nn_idx = tf_util.knn(adj_matrix, k=k)
edge_feature = tf_util.get_edge_feature(net, nn_idx=nn_idx, k=k)
net = tf_util.conv2d(edge_feature,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='dgcnn4',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=-2, keep_dims=True)
net_4 = net
#############################################################################
# aggregate block
#############################################################################
net = tf.concat([net_1, net_2, net_3, net_4], axis=-1)
net = tf_util.conv2d(net,
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay)
net = tf.reduce_max(net, axis=1, keep_dims=True)
# MLP on global point cloud vector
net = tf.reshape(net, [batch_size, -1])
net = tf_util.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net = tf_util.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net = tf_util.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
net = tf_util.fully_connected(net,
num_classes,
activation_fn=None,
scope='fc3')
return net, end_points
def get_model(point_cloud, input_label, is_training, cat_num, part_num, \
batch_size, num_point, weight_decay, bn_decay=None):
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)
k = 20
adj = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature,
is_training,
bn_decay,
K=3,
is_dist=True)
point_cloud_transformed = tf.matmul(point_cloud, transform)
input_image = tf.expand_dims(point_cloud_transformed, -1)
adj = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
out1 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv1',
bn_decay=bn_decay,
is_dist=True)
out2 = tf_util.conv2d(out1,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv2',
bn_decay=bn_decay,
is_dist=True)
net_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
adj = tf_util.pairwise_distance(net_1)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_1, nn_idx=nn_idx, k=k)
out3 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv3',
bn_decay=bn_decay,
is_dist=True)
out4 = tf_util.conv2d(out3,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv4',
bn_decay=bn_decay,
is_dist=True)
net_2 = tf.reduce_max(out4, axis=-2, keep_dims=True)
adj = tf_util.pairwise_distance(net_2)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_2, nn_idx=nn_idx, k=k)
out5 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv5',
bn_decay=bn_decay,
is_dist=True)
# out6 = tf_util.conv2d(out5, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv6', bn_decay=bn_decay, is_dist=True)
net_3 = tf.reduce_max(out5, axis=-2, keep_dims=True)
out7 = tf_util.conv2d(tf.concat([net_1, net_2, net_3], axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='adj_conv7',
bn_decay=bn_decay,
is_dist=True)
out_max = tf_util.max_pool2d(out7, [num_point, 1],
padding='VALID',
scope='maxpool')
one_hot_label_expand = tf.reshape(input_label, [batch_size, 1, 1, cat_num])
one_hot_label_expand = tf_util.conv2d(one_hot_label_expand,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='one_hot_label_expand',
bn_decay=bn_decay,
is_dist=True)
out_max = tf.concat(axis=3, values=[out_max, one_hot_label_expand])
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3, values=[expand, net_1, net_2, net_3])
net2 = tf_util.conv2d(concat,
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv1',
weight_decay=weight_decay,
is_dist=True)
net2 = tf_util.dropout(net2,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp1')
net2 = tf_util.conv2d(net2,
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv2',
weight_decay=weight_decay,
is_dist=True)
net2 = tf_util.dropout(net2,
keep_prob=0.6,
is_training=is_training,
scope='seg/dp2')
net2 = tf_util.conv2d(net2,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn_decay=bn_decay,
bn=True,
is_training=is_training,
scope='seg/conv3',
weight_decay=weight_decay,
is_dist=True)
net2 = tf_util.conv2d(net2,
part_num, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=None,
bn=False,
scope='seg/conv4',
weight_decay=weight_decay,
is_dist=True)
net2 = tf.reshape(net2, [batch_size, num_point, part_num])
return net2
def get_model(point_cloud,
is_training,
bn_decay=None,
weight_decay=None,
classes=13):
""" ConvNet baseline, input is BxNx9 gray image """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)
k = 20
adj = tf_util.pairwise_distance(point_cloud[:, :, 6:])
nn_idx = tf_util.knn(adj, k=k) # (batch, num_points, k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
out1 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv1',
bn_decay=bn_decay,
is_dist=True)
out2 = tf_util.conv2d(out1,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv2',
bn_decay=bn_decay,
is_dist=True)
net_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
adj = tf_util.pairwise_distance(net_1)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_1, nn_idx=nn_idx, k=k)
out3 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv3',
bn_decay=bn_decay,
is_dist=True)
out4 = tf_util.conv2d(out3,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv4',
bn_decay=bn_decay,
is_dist=True)
net_2 = tf.reduce_max(out4, axis=-2, keep_dims=True)
adj = tf_util.pairwise_distance(net_2)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(net_2, nn_idx=nn_idx, k=k)
out5 = tf_util.conv2d(edge_feature,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
weight_decay=weight_decay,
scope='adj_conv5',
bn_decay=bn_decay,
is_dist=True)
# out6 = tf_util.conv2d(out5, 64, [1,1],
# padding='VALID', stride=[1,1],
# bn=True, is_training=is_training, weight_decay=weight_decay,
# scope='adj_conv6', bn_decay=bn_decay, is_dist=True)
net_3 = tf.reduce_max(out5, axis=-2, keep_dims=True)
out7 = tf_util.conv2d(tf.concat([net_1, net_2, net_3], axis=-1),
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='adj_conv7',
bn_decay=bn_decay,
is_dist=True)
out_max = tf_util.max_pool2d(out7, [num_point, 1],
padding='VALID',
scope='maxpool')
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3, values=[expand, net_1, net_2, net_3])
# CONV
net = tf_util.conv2d(concat,
512, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='seg/conv1',
is_dist=True)
net = tf_util.conv2d(net,
256, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='seg/conv2',
is_dist=True)
net = tf_util.dropout(net,
keep_prob=0.7,
is_training=is_training,
scope='dp1')
net = tf_util.conv2d(net,
classes, [1, 1],
padding='VALID',
stride=[1, 1],
activation_fn=None,
scope='seg/conv3',
is_dist=True)
net = tf.squeeze(net, [2])
return net
def get_model(point_cloud, is_training, bn_decay=None):
""" Classification PointNet, input is BxNx3, output Bx40 """
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
end_points = {}
k_neighbors = (10, 20, 30, 40)
out_dims2neighbors = (8, 16, 16, 24)
n_heads = 4
origin_adj_matrix = tf_util.pairwise_distance(point_cloud)
nn_idx1 = tf_util.knn(origin_adj_matrix, k=k_neighbors[0])
nn_idx2 = tf_util.knn(origin_adj_matrix, k=k_neighbors[1])
nn_idx3 = tf_util.knn(origin_adj_matrix, k=k_neighbors[2])
nn_idx4 = tf_util.knn(origin_adj_matrix, k=k_neighbors[3])
# 对齐后的点云进入laplace—smooth过程
flag1 = 1
neighbors_edge_feature_list, local_feature_list = multiscale_view_filedGAP(
point_cloud,
nn_idx1=nn_idx1,
nn_idx2=nn_idx2,
nn_idx3=nn_idx3,
nn_idx4=nn_idx4,
out_dims2neighbors=out_dims2neighbors,
k_neighbors=k_neighbors,
multi_heads=n_heads,
scope_flag=flag1,
is_training=is_training,
bn_decay=bn_decay)
# 类似多通道注意力机制最后的输出，将得到的多通道数据串联起来 32*4。由于每个通道都最后做了 relu 激活，所以都为正值
neighbors_edge_feature = tf.concat(neighbors_edge_feature_list, axis=-1)
conv_net = tf_util.conv2d(neighbors_edge_feature,
128, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='conv1',
bn_decay=bn_decay)
net1 = conv_net # 128
conv_net = tf_util.conv2d(conv_net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='conv2',
bn_decay=bn_decay)
net2 = conv_net # 64
conv_net = tf_util.conv2d(conv_net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='conv3',
bn_decay=bn_decay)
net3 = conv_net # 64
conv_net = tf_util.conv2d(conv_net,
64, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='conv4',
bn_decay=bn_decay)
net4 = conv_net # 64
local_feature = tf.concat(local_feature_list, axis=-1)
net_out_concat = tf.concat([net1, net2, net3, net4, local_feature],
axis=-1)
net_out = tf_util.conv2d(net_out_concat,
1024, [1, 1],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='agg',
bn_decay=bn_decay)
net_out = tf.reduce_max(net_out, axis=1, keep_dims=True)
# MLP on global point cloud vector
net_out = tf.reshape(net_out, [batch_size, -1])
net_out = tf_util.fully_connected(net_out,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net_out = tf_util.dropout(net_out,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net_out = tf_util.fully_connected(net_out,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net_out = tf_util.dropout(net_out,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
net_out = tf_util.fully_connected(net_out,
40,
activation_fn=None,
scope='fc3')
return net_out, end_points
def get_model(point_cloud, input_label, is_training, cat_num, part_num, \
batch_size, num_point, weight_decay, bn_decay=None):
batch_size = point_cloud.get_shape()[0].value
num_point = point_cloud.get_shape()[1].value
input_image = tf.expand_dims(point_cloud, -1)
end_points = {}
k = 25
adj = tf_util.pairwise_distance(point_cloud)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
with tf.variable_scope('transform_net1') as sc:
transform = input_transform_net(edge_feature, is_training, bn_decay, K=3, is_dist=True)
point_cloud_transformed = tf.matmul(point_cloud, transform)
input_image = tf.expand_dims(point_cloud_transformed, -1)
adj = tf_util.pairwise_distance(point_cloud_transformed)
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(input_image, nn_idx=nn_idx, k=k)
out1 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv1', bn_decay=bn_decay, is_dist=True)
out2 = tf_util.conv2d(out1, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv2', bn_decay=bn_decay, is_dist=True)
net_max_1 = tf.reduce_max(out2, axis=-2, keep_dims=True)
net_mean_1 = tf.reduce_mean(out2, axis=-2, keep_dims=True)
out3 = tf_util.conv2d(tf.concat([net_max_1, net_mean_1], axis=-1), 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv3', bn_decay=bn_decay, is_dist=True)
adj = tf_util.pairwise_distance(tf.squeeze(out3, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(out3, nn_idx=nn_idx, k=k)
out4 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv4', bn_decay=bn_decay, is_dist=True)
net_max_2 = tf.reduce_max(out4, axis=-2, keep_dims=True)
net_mean_2 = tf.reduce_mean(out4, axis=-2, keep_dims=True)
out5 = tf_util.conv2d(tf.concat([net_max_2, net_mean_2], axis=-1), 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv5', bn_decay=bn_decay, is_dist=True)
adj = tf_util.pairwise_distance(tf.squeeze(out5, axis=-2))
nn_idx = tf_util.knn(adj, k=k)
edge_feature = tf_util.get_edge_feature(out5, nn_idx=nn_idx, k=k)
out6 = tf_util.conv2d(edge_feature, 64, [1,1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training, weight_decay=weight_decay,
scope='adj_conv6', bn_decay=bn_decay, is_dist=True)
net_max_3 = tf.reduce_max(out6, axis=-2, keep_dims=True)
net_mean_3 = tf.reduce_mean(out6, axis=-2, keep_dims=True)
out7 = tf_util.conv2d(tf.concat([out3, out5, out6], axis=-1), 1024, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='adj_conv13', bn_decay=bn_decay, is_dist=True)
out_max = tf_util.max_pool2d(out7, [num_point, 1], padding='VALID', scope='maxpool')
one_hot_label_expand = tf.reshape(input_label, [batch_size, 1, 1, cat_num])
one_hot_label_expand = tf_util.conv2d(one_hot_label_expand, 64, [1, 1],
padding='VALID', stride=[1,1],
bn=True, is_training=is_training,
scope='one_hot_label_expand', bn_decay=bn_decay, is_dist=True)
out_max = tf.concat(axis=3, values=[out_max, one_hot_label_expand])
expand = tf.tile(out_max, [1, num_point, 1, 1])
concat = tf.concat(axis=3, values=[expand,
net_max_1,
net_mean_1,
out3,
net_max_2,
net_mean_2,
out5,
net_max_3,
net_mean_3,
out6,
out7])
net2 = tf_util.conv2d(concat, 256, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
bn=True, is_training=is_training, scope='seg/conv1', weight_decay=weight_decay, is_dist=True)
net2 = tf_util.dropout(net2, keep_prob=0.6, is_training=is_training, scope='seg/dp1')
net2 = tf_util.conv2d(net2, 256, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
bn=True, is_training=is_training, scope='seg/conv2', weight_decay=weight_decay, is_dist=True)
net2 = tf_util.dropout(net2, keep_prob=0.6, is_training=is_training, scope='seg/dp2')
net2 = tf_util.conv2d(net2, 128, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
bn=True, is_training=is_training, scope='seg/conv3', weight_decay=weight_decay, is_dist=True)
net2 = tf_util.conv2d(net2, 128, [1,1], padding='VALID', stride=[1,1], bn_decay=bn_decay,
bn=True, is_training=is_training, scope='seg/conv4', weight_decay=weight_decay, is_dist=True)
net2 = tf_util.conv2d(net2, part_num, [1,1], padding='VALID', stride=[1,1], activation_fn=None,
bn=False, scope='seg/conv5', weight_decay=weight_decay, is_dist=True)
net2 = tf.reshape(net2, [batch_size, num_point, part_num])
return net2
def get_loss(seg_pred, seg):
per_instance_seg_loss = tf.reduce_mean(tf.nn.sparse_softmax_cross_entropy_with_logits(logits=seg_pred, labels=seg), axis=1)
seg_loss = tf.reduce_mean(per_instance_seg_loss)
per_instance_seg_pred_res = tf.argmax(seg_pred, 2)
return seg_loss, per_instance_seg_loss, per_instance_seg_pred_res