def wnet_model(img,is_training,class_num,bn_decay=None):
with tf.variable_scope('input_layer'):
out0=input_down_size_unit(img,64,192,is_training,bn_decay,'input_pre')
skip0=out0
with tf.variable_scope('intermidate_layer'):
out1 ,skip1= skip_sbstract_layer(skip0,out0,[16,16,32],[[5,5],[3,3],[1,1]],is_training,bn_decay=bn_decay,scope='inception_1')
out2 ,skip2= skip_sbstract_layer(skip1,out1,[32,32,64],[[5,5],[3,3],[1,1]],is_training,bn_decay=bn_decay,scope='inception_2')
out3 ,skip3= skip_sbstract_layer(skip2,out2,[64,64,96],[[5,5],[3,3],[1,1]],is_training,bn_decay=bn_decay,scope='inception_3',maxpool=False)
out = output_down_size_unit(tf.concat((out3,skip3),-1),256,is_training,bn_decay,'out_downsize')
with tf.variable_scope('output_layer'):
out = basic_tf.fully_connected(out,1024, bn=True, is_training=is_training, scope='fc1', bn_decay=bn_decay)
out = basic_tf.dropout(out, keep_prob=0.5, is_training=is_training, scope='dp1')
out = basic_tf.fully_connected(out,128, bn=True, is_training=is_training, scope='fc2', bn_decay=bn_decay)
out = basic_tf.dropout(out, keep_prob=0.5, is_training=is_training, scope='dp2')
out = basic_tf.fully_connected(out,class_num, bn=True, is_training=is_training, scope='fc3', activation_fn=None,bn_decay=bn_decay)
out_pred = tf.argmax(out,1,name="final")
return out,out_pred
def fullyconnected_layer(input, batch_size, class_num, scope, is_training,
bn_decay):
with tf.variable_scope(scope):
net = tf.reshape(input, (batch_size, -1))
net = basic_tf.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='fc1',
bn_decay=bn_decay)
net = basic_tf.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp1')
net = basic_tf.fully_connected(net,
128,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
net = basic_tf.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
net = basic_tf.fully_connected(net,
class_num,
activation_fn=None,
scope='fc3')
return net
def basic_detectModel(img, is_training, bn_decay, cell_size, num_class,
box_num): #512->4
with tf.variable_scope('conv_unit1'):
out = basic_tf.conv2d(img, 16, [3, 3], 'conv_11', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_11', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 32, [3, 3], 'conv_12', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_12', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 64, [3, 3], 'conv_13', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_13', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 128, [3, 3], 'conv_14', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_14', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 256, [3, 3], 'conv_15', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_15', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 512, [3, 3], 'conv_16', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_16', [2, 2], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_17', [2, 2], 'SAME')
with tf.variable_scope('conv_unit2'):
out1 = basic_tf.conv2d(out, 1024, [3, 3], 'conv_21', [1, 1], 'SAME')
out1 = basic_tf.conv2d(out1, 512, [1, 1], 'conv_22', [1, 1], 'SAME')
with tf.variable_scope('fully_connected_unit'):
out2 = tf.reshape(out1, (int(out1._shape[0]), -1)) #b,8192
out2 = basic_tf.fully_connected(out2, 4096, 'fc1')
out2 = basic_tf.dropout(out2, is_training, 'dp1', 0.5)
out2 = basic_tf.fully_connected(out2, 1024, 'fc2')
out2 = basic_tf.dropout(out2, is_training, 'dp2', 0.5)
out2 = basic_tf.fully_connected(
out2, cell_size[0] * cell_size[1] * (class_num + box_num * 5),
'fc3')
with tf.variable_scope('output_unit'):
n1 = cell_size[0] * cell_size[1] * num_class #4*4*2=32
n2 = n1 + cell_size[0] * cell_size[1] * box_num #32+4*4*2=64
class_pred = tf.reshape(
out2[:, 0:n1],
(-1, cell_size[0], cell_size[1], num_class)) #(b,4,4,2)
scales = tf.reshape(
out2[:,
n1:n2], (-1, cell_size[0], cell_size[1], box_num)) #(b,4,4,2)
boxes = tf.reshape(
out2[:, n2:],
(-1, cell_size[0], cell_size[1], box_num * 4)) #(b,4,4,8)
pred = tf.concat([class_pred, scales, boxes], 3) #(b,4,4,12)
return pred
def get_model_single_seg(point_cloud, is_training, bn_decay=None,is_tnet=False):
b = point_cloud.get_shape()[0].value
n = point_cloud.get_shape()[1].value
end_points = {}
l0_xyz = tf.slice(point_cloud, [0,0,0], [-1,-1,3])
l0_points = point_cloud
if is_tnet:
tnet_spec={'mlp':[64,128,1024], 'is_training':is_training, 'bn_decay':bn_decay}
else:
tnet_spec=None
end_points['l0_xyz'] =l0_xyz
#Abstraction layers
l1_xyz, l1_points,_ = pointnet_AB_module(l0_xyz, l0_points, m=51200, r=0.4, ns=64, mlp=[32,64,128], mlp2=None, group_all=False, is_training=is_training, bn_decay=bn_decay, scope='layer1',tnet_spec=tnet_spec)
l2_xyz, l2_points,_= pointnet_AB_module(l1_xyz, l1_points,m=12800, r=1.6, ns=128, mlp=[64,128,256], mlp2=None, group_all=False, is_training=is_training, bn_decay=bn_decay, scope='layer2',tnet_spec=tnet_spec)
l3_xyz, l3_points,_ = pointnet_AB_module(l2_xyz, l2_points, m=5120, r=3.2, ns=64, mlp=[256,256,512], mlp2=None, group_all=False, is_training=is_training, bn_decay=bn_decay, scope='layer3',tnet_spec=tnet_spec)
l4_xyz, l4_points,_= pointnet_AB_module( l3_xyz, l3_points,m=2048, r=6.4, ns=64, mlp=[256,512,512], mlp2=None, group_all=False, is_training=is_training, bn_decay=bn_decay, scope='layer4',tnet_spec=tnet_spec)
l5_xyz, l5_points,_ = pointnet_AB_module(l4_xyz, l4_points, m=1024, r=19.2, ns=64, mlp=[512,1024,1024], mlp2=None, group_all=False, is_training=is_training, bn_decay=bn_decay, scope='layer5',tnet_spec=tnet_spec)
l6_xyz, l6_points,_= pointnet_AB_module(l5_xyz, l5_points,m=512, r=38.4, ns=64, mlp=[1024,1024,2048], mlp2=None, group_all=False, is_training=is_training, bn_decay=bn_decay, scope='layer6',tnet_spec=tnet_spec)
l7_xyz, l7_points,_ = pointnet_AB_module(l6_xyz, l6_points, m=None, r=None, ns=None, mlp=[2048,2048,4096], mlp2=None, group_all=True, is_training=is_training, bn_decay=bn_decay, scope='layer7')
# Fully connected layers
netc = tf.reshape(l7_points, [b,-1])
netc = basic_tf.fully_connected(netc,2048, bn=True, is_training=is_training, scope='fc1', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.8, is_training=is_training, scope='dp1')
netc = basic_tf.fully_connected(netc,1024, bn=True, is_training=is_training, scope='fc2', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.8, is_training=is_training, scope='dp2')
netc = basic_tf.fully_connected(netc, 512, bn=True, is_training=is_training, scope='fc3', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.5, is_training=is_training, scope='dp3')
netc = basic_tf.fully_connected(netc, 128, bn=True, is_training=is_training, scope='fc4', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.5, is_training=is_training, scope='dp4')
netc = basic_tf.fully_connected(netc, 40, activation_fn=None, scope='fc5')
# Feature Propagation layers&fully connected layers for segmentation
l6_points = pointnet_fp_module(l6_xyz, l7_xyz, l6_points, l7_points, [2048,2048,1024], is_training, bn_decay, scope='fa_layer1')
l5_points = pointnet_fp_module(l5_xyz, l6_xyz, l5_points, l6_points, [1024,1024], is_training, bn_decay, scope='fa_layer2')
l4_points = pointnet_fp_module(l4_xyz, l5_xyz, l4_points, l5_points, [1024,512], is_training, bn_decay, scope='fa_layer3')
l3_points = pointnet_fp_module(l3_xyz, l4_xyz, l3_points, l4_points, [512,512], is_training, bn_decay, scope='fa_layer4')
l2_points = pointnet_fp_module(l2_xyz, l3_xyz, l2_points, l3_points, [512,256], is_training, bn_decay, scope='fa_layer5')
l1_points = pointnet_fp_module(l1_xyz, l2_xyz, l1_points, l2_points, [256,128], is_training, bn_decay, scope='fa_layer6')
l0_points = pointnet_fp_module(l0_xyz, l1_xyz, l0_points, l1_points, [128,128,128], is_training, bn_decay, scope='fa_layer7') #(32,102400,128)
net = basic_tf.conv1d(l0_points, 128, 1, padding='VALID', bn=True, is_training=is_training, scope='fcs1', bn_decay=bn_decay)
end_points['feats'] = net
net = basic_tf.dropout(net, keep_prob=0.5, is_training=is_training, scope='dps1')
net = basic_tf.conv1d(net, 50, 1, padding='VALID', activation_fn=None, scope='fcs2')
return netc,net,end_points
def get_model_multi_seg(point_cloud, is_training, bn_decay=None):
b = point_cloud.get_shape()[0].value
n = point_cloud.get_shape()[1].value
c=point_cloud.get_shape()[2].value
end_points = {}
l0_points = point_cloud
if c>3:
l0_xyz = tf.slice(point_cloud,[0,0,0],[-1,-1,3])
else:
l0_xyz = point_cloud
#Abstraction layers
l1_xyz, l1_points = pointnet_AB_module_msg(l0_xyz, l0_points, 51200, [0.2,0.4,0.8], [32,64,96], [[32,32,64], [64,64,128], [64,96,128]], is_training, bn_decay, scope='layer1')
l2_xyz, l2_points = pointnet_AB_module_msg(l1_xyz, l1_points, 12800, [0.8,1.6,3.2], [128,128,256], [[64,64,128], [128,128,256], [128,128,256]], is_training, bn_decay, scope='layer2')
l3_xyz, l3_points = pointnet_AB_module_msg(l2_xyz, l2_points, 5120, [1.6,3.2,6.4], [64,128,128], [[128,128,256], [256,256,512], [512,512,1024]], is_training, bn_decay, scope='layer3')
l4_xyz, l4_points = pointnet_AB_module_msg(l3_xyz, l3_points,2048, [3.2,6.4,12.8], [128,128,256], [[256,256,512], [512,512,1024], [1024,1024,2048]], is_training, bn_decay, scope='layer4')
l5_xyz, l5_points = pointnet_AB_module_msg(l4_xyz, l4_points,1024, [12.8,25.6], [64,128], [[512,1024,1024], [1024,2048,2048]], is_training, bn_decay, scope='layer5')
l6_xyz, l6_points = pointnet_AB_module_msg(l5_xyz, l5_points,512, [25.6,51.2], [32,64], [[512,1024,1024], [1024,2048,2048]], is_training, bn_decay, scope='layer6')
l7_xyz, l7_points, _ = pointnet_AB_module(l6_xyz, l6_points,m=None, r=None, ns=None, mlp=[2048,2048,4096], mlp2=None, group_all=True, is_training=is_training, bn_decay=bn_decay, scope='layer7')
# Fully connected layers for classification
netc = tf.reshape(l7_points, [b, -1])
netc = basic_tf.fully_connected(netc,2048, bn=True, is_training=is_training, scope='fc1', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.8, is_training=is_training, scope='dp1')
netc = basic_tf.fully_connected(netc, 1024, bn=True, is_training=is_training, scope='fc2', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.8, is_training=is_training, scope='dp2')
netc = basic_tf.fully_connected(netc, 512, bn=True, is_training=is_training, scope='fc3', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.5, is_training=is_training, scope='dp3')
netc = basic_tf.fully_connected(netc, 128, bn=True, is_training=is_training, scope='fc4', bn_decay=bn_decay)
netc = basic_tf.dropout(netc, keep_prob=0.5, is_training=is_training, scope='dp4')
netc = basic_tf.fully_connected(netc, 40, activation_fn=None, scope='fc5')
# Feature Propagation layers&fully connected layers for segmentation
l6_points = pointnet_fp_module(l6_xyz, l7_xyz, l6_points, l7_points, [2048,2048,1024], is_training, bn_decay, scope='fa_layer1')
l5_points = pointnet_fp_module(l5_xyz, l6_xyz, l5_points, l6_points, [1024,1024], is_training, bn_decay, scope='fa_layer2')
l4_points = pointnet_fp_module(l4_xyz, l5_xyz, l4_points, l5_points, [1024,512], is_training, bn_decay, scope='fa_layer3')
l3_points = pointnet_fp_module(l3_xyz, l4_xyz, l3_points, l4_points, [512,512], is_training, bn_decay, scope='fa_layer4')
l2_points = pointnet_fp_module(l2_xyz, l3_xyz, l2_points, l3_points, [512,256], is_training, bn_decay, scope='fa_layer5')
l1_points = pointnet_fp_module(l1_xyz, l2_xyz, l1_points, l2_points, [256,256], is_training, bn_decay, scope='fa_layer6')
l0_points = pointnet_fp_module(l0_xyz, l1_xyz, l0_points, l1_points, [256,128,128], is_training, bn_decay, scope='fa_layer7') #(32,102400,128)
net = basic_tf.conv1d(l0_points,128, 1, padding='VALID', bn=True, is_training=is_training, scope='fcs1', bn_decay=bn_decay)
end_points['feats'] = net
net = basic_tf.dropout(net, keep_prob=0.5, is_training=is_training, scope='dps1')
net = basic_tf.conv1d(net, 50, 1, padding='VALID', activation_fn=None, scope='fcs3')
return netc,net,end_points
def feature_transform_net(inputs, mlp, is_training, bn_decay=None, K=64):
""" Feature Transform Net, input is BxNx1xK
mlp:list of output_channels
Return:
transformed_inputs (b,n,k)"""
b = xyz.get_shape()[0].value
n = xyz.get_shape()[1].value
net = inputs
for i, num_out_channel in enumerate(mlp):
net = basic_tf.conv2d(net,
num_out_channel, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn,
is_training=is_training,
scope='tconv%d' % (i + 1),
bn_decay=bn_decay)
net = basic_tf.max_pool2d(net, [n, 1], padding='VALID', scope='tmaxpool')
net = tf.reshape(net, [b, -1])
net = basic_tf.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='tfc1',
bn_decay=bn_decay)
net = basic_tf.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='tfc2',
bn_decay=bn_decay)
with tf.variable_scope('transform_feat') as sc:
weights = tf.get_variable('weights', [256, K * K],
initializer=tf.constant_initializer(0.0),
dtype=tf.float32)
biases = tf.get_variable('biases', [K * K],
initializer=tf.constant_initializer(0.0),
dtype=tf.float32)
biases += tf.constant(np.eye(K).flatten(), dtype=tf.float32)
transform = tf.matmul(net, weights)
transform = tf.nn.bias_add(transform, biases)
transform = tf.reshape(transform, [b, K, K])
transformed_inputs = tf.matmul(input, transform)
return transformed_inputs
def basic_detectModel(img, is_training, bn_decay, num_class): #512->4
with tf.variable_scope('conv_unit1_G'):
out = basic_tf.conv2d(img, 16, [3, 3], 'conv_11', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_11', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 32, [3, 3], 'conv_12', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_12', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 64, [3, 3], 'conv_13', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_13', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 128, [3, 3], 'conv_14', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_14', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 256, [3, 3], 'conv_15', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_15', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 512, [3, 3], 'conv_16', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_16', [2, 2], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_17', [2, 2], 'SAME')
with tf.variable_scope('conv_unit2'):
out1 = basic_tf.conv2d(out, 1024, [3, 3], 'conv_21', [1, 1], 'SAME')
out1 = basic_tf.conv2d(out1, 512, [1, 1], 'conv_22', [1, 1], 'SAME')
out1 = basic_tf.avg_pool2d(out1, [2, 2], 'pre_avepool', [2, 2], 'SAME')
with tf.variable_scope('fully_connected_unit_G'):
out2 = tf.reshape(out1, (int(out1._shape[0]), -1)) #b,8192
#out2 = basic_tf.fully_connected(out2,4096,'fc1')
#out2 = basic_tf.dropout(out2,is_training,'dp1',0.5)
out2 = basic_tf.fully_connected(out2, 1024, 'fc2')
out2 = basic_tf.dropout(out2, is_training, 'dp2', 0.5)
out2 = basic_tf.fully_connected(out2, 128, 'fc3')
out2 = basic_tf.dropout(out2, is_training, 'dp3', 0.5)
with tf.variable_scope('output_unit_G'):
pred = basic_tf.fully_connected(out2, (num_class + 4),
'fc4',
activation_fn=None)
return pred
def brute_classify(img, num_class, is_training, bn_decay): #512->4
with tf.variable_scope('conv_unit1'):
out = basic_tf.conv2d(img, 16, [3, 3], 'conv_11', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_11', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 32, [3, 3], 'conv_12', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_12', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 64, [3, 3], 'conv_13', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_13', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 128, [3, 3], 'conv_14', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_14', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 256, [3, 3], 'conv_15', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_15', [2, 2], 'SAME')
out = basic_tf.conv2d(out, 512, [3, 3], 'conv_16', [1, 1], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_16', [2, 2], 'SAME')
out = basic_tf.max_pool2d(out, [2, 2], 'maxpool_17', [2, 2], 'SAME')
with tf.variable_scope('conv_unit2'):
out1 = basic_tf.conv2d(out, 1024, [3, 3], 'conv_21', [1, 1], 'SAME')
out1 = basic_tf.conv2d(out1, 512, [1, 1], 'conv_22', [1, 1], 'SAME')
out1 = basic_tf.avg_pool2d(out1, [2, 2], 'pre_avepool', [2, 2], 'SAME')
with tf.variable_scope('fully_connected_unit'):
out2 = tf.reshape(out1, (int(out1._shape[0]), -1)) #b,4096
out2 = basic_tf.fully_connected(out2, 1024, 'fc1')
out2 = basic_tf.dropout(out2, is_training, 'dp1', 0.5)
out2 = basic_tf.fully_connected(out2, 128, 'fc2')
out2 = basic_tf.dropout(out2, is_training, 'dp2', 0.5)
pred = basic_tf.fully_connected(out2, num_class, 'fc3')
print(pred)
return pred
def input_transform_net(xyz, mlp, is_training, bn_decay=None, K=3):
""" Input (XYZ) Transform Net, input is BxNx3 gray image
mlp:list of output_channels
Return:
transformed_xyz (b,n,3)"""
b = xyz.get_shape()[0].value
n = xyz.get_shape()[1].value
print('transing')
input = tf.expand_dims(xyz, -1) #(b,n,3,1)
for i, num_out_channel in enumerate(mlp):
if i == 0:
net = basic_tf.conv2d(input,
num_out_channel, [1, 3],
padding='VALID',
stride=[1, 1],
bn=True,
is_training=is_training,
scope='tconv%d' % (i + 1),
bn_decay=bn_decay)
else:
net = basic_tf.conv2d(net,
num_out_channel, [1, 1],
padding='VALID',
stride=[1, 1],
bn=bn_decay,
is_training=is_training,
scope='tconv%d' % (i + 1),
bn_decay=bn_decay) #(b,n,mlp[-1])
net = basic_tf.max_pool2d(net, [n, 1], padding='VALID',
scope='tmaxpool') #(b,1,mlp[-1])
net = tf.reshape(net, [b, -1]) #(b * mlp[-1])
net = basic_tf.fully_connected(net,
512,
bn=True,
is_training=is_training,
scope='tfc1',
bn_decay=bn_decay) #(b,512)
net = basic_tf.fully_connected(net,
256,
bn=True,
is_training=is_training,
scope='tfc2',
bn_decay=bn_decay) #(b,256)
print('transing2')
with tf.variable_scope('transform_XYZ') as sc:
assert (K == 3)
weights = tf.get_variable('weights', [256, 3 * K],
initializer=tf.constant_initializer(0.0),
dtype=tf.float32)
biases = tf.get_variable('biases', [3 * K],
initializer=tf.constant_initializer(0.0),
dtype=tf.float32)
biases += tf.constant([1, 0, 0, 0, 1, 0, 0, 0, 1], dtype=tf.float32)
transform = tf.matmul(net, weights)
transform = tf.nn.bias_add(transform, biases) #now,net_shape:(b,3*K)
transform = tf.reshape(transform, [b, 3, K])
transformed_xyz = tf.matmul(xyz, transform)
return transformed_xyz
def color_net(rgb, is_training, bn_decay=None):
with tf.variable_scope('input_layer'):
h = rgb.get_shape()[1].value
w = rgb.get_shape()[2].value
b = rgb.get_shape()[0].value
og = [h, w]
end_data = {}
end_data['rgb_data'] = rgb
#end_data['srgb_data']=srgb #both of them are been normalized
out1 = basic_tf.conv2d(rgb, 96, [1, 1], 'input_conv', [1, 1], 'SAME')
out1 = basic_tf.max_pool2d(out1, [2, 2], 'input_pool', [1, 1], 'SAME')
with tf.variable_scope('intermidate_layer'):
for i, kernels in enumerate(list_of_kernel):
mlps = list_of_mlplist[i]
out1 = bm.ssc_color_info_abstraction(out1,
mlps,
is_training=is_training,
bn_decay=bn_decay,
scope='ssc_section_%d' % (i),
kernel_size=kernels,
bn=True)
if i == 0:
hyper_colume = out1
else:
hyper_colume = tf.concat([hyper_colume, out1], -1)
hyper_colume = basic_tf.avg_pool2d(hyper_colume, [2, 2],
'medium_avepool', [1, 1], 'SAME')
c = hyper_colume.get_shape()[-1].value
print(hyper_colume.shape)
hyper_colume = tf.reshape(hyper_colume, (b * h * w, c))
with tf.variable_scope('output_layer'):
out = basic_tf.fully_connected(hyper_colume,
256,
bn=True,
is_training=is_training,
scope='fc2',
bn_decay=bn_decay)
out = basic_tf.dropout(out,
keep_prob=0.5,
is_training=is_training,
scope='dp2')
out = basic_tf.fully_connected(out,
64,
bn=True,
is_training=is_training,
scope='fc3',
bn_decay=bn_decay)
out = basic_tf.dropout(out,
keep_prob=0.5,
is_training=is_training,
scope='dp3')
out = basic_tf.fully_connected(out,
3,
bn=True,
is_training=is_training,
scope='fc4',
bn_decay=bn_decay)
pred = tf.reshape(out, (b, h, w, 3))
return pred, end_data
