def build_model():
#################
# Regular model #
#################
input_size = data_sizes["sliced:data:singleslice:4ch"]
l0 = nn.layers.InputLayer(input_size)
l1a = nn.layers.dnn.Conv2DDNNLayer(l0, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=64, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l1b = nn.layers.dnn.Conv2DDNNLayer(l1a, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=64, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l1 = nn.layers.dnn.MaxPool2DDNNLayer(l1b, pool_size=(2,2), stride=(2,2))
l2a = nn.layers.dnn.Conv2DDNNLayer(l1, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=128, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l2b = nn.layers.dnn.Conv2DDNNLayer(l2a, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=128, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l2 = nn.layers.dnn.MaxPool2DDNNLayer(l2b, pool_size=(2,2), stride=(2,2))
l3a = nn.layers.dnn.Conv2DDNNLayer(l2, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=256, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l3b = nn.layers.dnn.Conv2DDNNLayer(l3a, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=256, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l3c = nn.layers.dnn.Conv2DDNNLayer(l3b, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=256, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l3 = nn.layers.dnn.MaxPool2DDNNLayer(l3c, pool_size=(2,2), stride=(2,2))
l4a = nn.layers.dnn.Conv2DDNNLayer(l3, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=512, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l4b = nn.layers.dnn.Conv2DDNNLayer(l4a, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=512, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l4c = nn.layers.dnn.Conv2DDNNLayer(l4b, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=512, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l4 = nn.layers.dnn.MaxPool2DDNNLayer(l4c, pool_size=(2,2), stride=(2,2))
l5a = nn.layers.dnn.Conv2DDNNLayer(l4, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=512, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l5b = nn.layers.dnn.Conv2DDNNLayer(l5a, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=512, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l5c = nn.layers.dnn.Conv2DDNNLayer(l5b, W=nn.init.Orthogonal("relu"), filter_size=(3,3), num_filters=512, stride=(1,1), pad="same", nonlinearity=nn.nonlinearities.rectify)
l5 = nn.layers.dnn.MaxPool2DDNNLayer(l5c, pool_size=(2,2), stride=(2,2))
# Systole Dense layers
ldsys1 = nn.layers.DenseLayer(l5, num_units=512, W=nn.init.Orthogonal("relu"), b=nn.init.Constant(0.1), nonlinearity=nn.nonlinearities.rectify)
ldsys1drop = nn.layers.dropout(ldsys1, p=0.5)
ldsys2 = nn.layers.DenseLayer(ldsys1drop, num_units=512, W=nn.init.Orthogonal("relu"),b=nn.init.Constant(0.1), nonlinearity=nn.nonlinearities.rectify)
ldsys2drop = nn.layers.dropout(ldsys2, p=0.5)
ldsys3 = nn.layers.DenseLayer(ldsys2drop, num_units=600, W=nn.init.Orthogonal("relu"), b=nn.init.Constant(0.1), nonlinearity=nn.nonlinearities.softmax)
ldsys3drop = nn.layers.dropout(ldsys3, p=0.5) # dropout at the output might encourage adjacent neurons to correllate
ldsys3dropnorm = layers.NormalisationLayer(ldsys3drop)
l_systole = layers.CumSumLayer(ldsys3dropnorm)
# Diastole Dense layers
lddia1 = nn.layers.DenseLayer(l5, num_units=512, W=nn.init.Orthogonal("relu"), b=nn.init.Constant(0.1), nonlinearity=nn.nonlinearities.rectify)
lddia1drop = nn.layers.dropout(lddia1, p=0.5)
lddia2 = nn.layers.DenseLayer(lddia1drop, num_units=512, W=nn.init.Orthogonal("relu"),b=nn.init.Constant(0.1), nonlinearity=nn.nonlinearities.rectify)
lddia2drop = nn.layers.dropout(lddia2, p=0.5)
lddia3 = nn.layers.DenseLayer(lddia2drop, num_units=600, W=nn.init.Orthogonal("relu"), b=nn.init.Constant(0.1), nonlinearity=nn.nonlinearities.softmax)
lddia3drop = nn.layers.dropout(lddia3, p=0.5) # dropout at the output might encourage adjacent neurons to correllate
lddia3dropnorm = layers.NormalisationLayer(lddia3drop)
l_diastole = layers.CumSumLayer(lddia3dropnorm)
return {
"inputs":{
"sliced:data:singleslice:4ch": l0
},
"outputs": {
"systole": l_systole,
"diastole": l_diastole,
},
"regularizable": {
ldsys1: l2_weight,
ldsys2: l2_weight,
ldsys3: l2_weight_out,
lddia1: l2_weight,
lddia2: l2_weight,
lddia3: l2_weight_out,
},
}
def build_model(input_layer=None):
#################
# Regular model #
#################
input_size = data_sizes["sliced:data:singleslice"]
if input_layer:
l0 = input_layer
else:
l0 = nn.layers.InputLayer(input_size)
l1 = jonas_residual(l0,
num_filters=64,
num_conv=2,
filter_size=(3, 3),
pool_size=(2, 2))
l2 = jonas_residual(l1,
num_filters=128,
num_conv=2,
filter_size=(3, 3),
pool_size=(2, 2))
l3 = jonas_residual(l2,
num_filters=256,
num_conv=3,
filter_size=(3, 3),
pool_size=(2, 2))
l4 = jonas_residual(l3,
num_filters=512,
num_conv=3,
filter_size=(3, 3),
pool_size=(2, 2))
l5 = jonas_residual(l4,
num_filters=512,
num_conv=3,
filter_size=(3, 3),
pool_size=(2, 2))
# Systole Dense layers
ldsys1 = nn.layers.DenseLayer(l5,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
ldsys1drop = nn.layers.dropout(ldsys1, p=0.5)
ldsys2 = nn.layers.DenseLayer(ldsys1drop,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
ldsys2drop = nn.layers.dropout(ldsys2, p=0.5)
ldsys3 = nn.layers.DenseLayer(ldsys2drop,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
ldsys3drop = nn.layers.dropout(
ldsys3, p=0.5
) # dropout at the output might encourage adjacent neurons to correllate
ldsys3dropnorm = layers.NormalisationLayer(ldsys3drop)
l_systole = layers.CumSumLayer(ldsys3dropnorm)
# Diastole Dense layers
lddia1 = nn.layers.DenseLayer(l5,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
lddia1drop = nn.layers.dropout(lddia1, p=0.5)
lddia2 = nn.layers.DenseLayer(lddia1drop,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
lddia2drop = nn.layers.dropout(lddia2, p=0.5)
lddia3 = nn.layers.DenseLayer(lddia2drop,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
lddia3drop = nn.layers.dropout(
lddia3, p=0.5
) # dropout at the output might encourage adjacent neurons to correllate
lddia3dropnorm = layers.NormalisationLayer(lddia3drop)
l_diastole = layers.CumSumLayer(lddia3dropnorm)
return {
"inputs": {
"sliced:data:singleslice": l0
},
"outputs": {
"systole": l_systole,
"diastole": l_diastole,
},
"regularizable": {
ldsys1: l2_weight,
ldsys2: l2_weight,
ldsys3: l2_weight_out,
lddia1: l2_weight,
lddia2: l2_weight,
lddia3: l2_weight_out,
},
"meta_outputs": {
"systole": ldsys2,
"diastole": lddia2,
}
}
def build_model():
#################
# Regular model #
#################
input_size = data_sizes["sliced:data:sax"]
input_size_mask = data_sizes["sliced:data:sax:is_not_padded"]
input_size_locations = data_sizes["sliced:data:sax:locations"]
l0 = nn.layers.InputLayer(input_size)
lin_slice_mask = nn.layers.InputLayer(input_size_mask)
lin_slice_locations = nn.layers.InputLayer(input_size_locations)
# PREPROCESS SLICES SEPERATELY
# Convolutional layers and some dense layers are defined in a submodel
l0_slices = nn.layers.ReshapeLayer(l0, (-1, [2], [3], [4]))
from . import je_ss_jonisc64small_360
submodel = je_ss_jonisc64small_360.build_model(l0_slices)
# Systole Dense layers
ldsysout = submodel["outputs"]["systole"]
# Diastole Dense layers
lddiaout = submodel["outputs"]["diastole"]
# AGGREGATE SLICES PER PATIENT
# Systole
ldsys_pat_in = nn.layers.ReshapeLayer(ldsysout, (-1, nr_slices, [1]))
ldsys_rnn = rnn_layer(ldsys_pat_in,
num_units=256,
mask_input=lin_slice_mask)
# ldsys_rnn_drop = nn.layers.dropout(ldsys_rnn, p=0.5)
ldsys3 = nn.layers.DenseLayer(ldsys_rnn,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
ldsys3drop = nn.layers.dropout(
ldsys3, p=0.5
) # dropout at the output might encourage adjacent neurons to correllate
ldsys3dropnorm = layers.NormalisationLayer(ldsys3drop)
l_systole = layers.CumSumLayer(ldsys3dropnorm)
# Diastole
lddia_pat_in = nn.layers.ReshapeLayer(lddiaout, (-1, nr_slices, [1]))
lddia_rnn = rnn_layer(lddia_pat_in,
num_units=256,
mask_input=lin_slice_mask)
# lddia_rnn_drop = nn.layers.dropout(lddia_rnn, p=0.5)
lddia3 = nn.layers.DenseLayer(lddia_rnn,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
lddia3drop = nn.layers.dropout(
lddia3, p=0.5
) # dropout at the output might encourage adjacent neurons to correllate
lddia3dropnorm = layers.NormalisationLayer(lddia3drop)
l_diastole = layers.CumSumLayer(lddia3dropnorm)
submodels = [submodel]
return {
"inputs": {
"sliced:data:sax": l0,
"sliced:data:sax:is_not_padded": lin_slice_mask,
"sliced:data:sax:locations": lin_slice_locations,
},
"outputs": {
"systole": l_systole,
"diastole": l_diastole,
},
"regularizable":
dict({
lddia3: l2_weight_out,
ldsys3: l2_weight_out,
}, **{
k: v
for d in [
model["regularizable"] for model in submodels
if "regularizable" in model
] for k, v in list(d.items())
}),
"pretrained": {
je_ss_jonisc64small_360.__name__: submodel["outputs"],
}
}
def build_model():
#################
# Regular model #
#################
input_size_age = data_sizes["sliced:meta:PatientAge"]
input_size_sex = data_sizes["sliced:meta:PatientSex"]
l0_age = nn.layers.InputLayer(input_size_age)
l0_sex = nn.layers.InputLayer(input_size_sex)
l0 = nn.layers.ConcatLayer([l0_age, l0_sex], axis=1)
# Systole Dense layers
ldsys1 = nn.layers.DenseLayer(l0,
num_units=128,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
ldsys1drop = nn.layers.dropout(ldsys1, p=0.5)
ldsys2 = nn.layers.DenseLayer(ldsys1drop,
num_units=128,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
ldsys2drop = nn.layers.dropout(ldsys2, p=0.5)
ldsys3 = nn.layers.DenseLayer(ldsys2drop,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
ldsys3drop = nn.layers.dropout(
ldsys3, p=0.9
) # dropout at the output might encourage adjacent neurons to correllate
ldsys3dropnorm = layers.NormalisationLayer(ldsys3drop)
l_systole = layers.CumSumLayer(ldsys3dropnorm)
# Diastole Dense layers
lddia1 = nn.layers.DenseLayer(l0,
num_units=128,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
lddia1drop = nn.layers.dropout(lddia1, p=0.5)
lddia2 = nn.layers.DenseLayer(lddia1drop,
num_units=128,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
lddia2drop = nn.layers.dropout(lddia2, p=0.5)
lddia3 = nn.layers.DenseLayer(lddia2drop,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
lddia3drop = nn.layers.dropout(
lddia3, p=0.9
) # dropout at the output might encourage adjacent neurons to correllate
lddia3dropnorm = layers.NormalisationLayer(lddia3drop)
l_diastole = layers.CumSumLayer(lddia3dropnorm)
return {
"inputs": {
"sliced:meta:PatientAge": l0_age,
"sliced:meta:PatientSex": l0_sex,
},
"outputs": {
"systole": l_systole,
"diastole": l_diastole,
},
"regularizable": {
ldsys1: l2_weight,
ldsys2: l2_weight,
ldsys3: l2_weight_out,
lddia1: l2_weight,
lddia2: l2_weight,
lddia3: l2_weight_out,
},
}
def build_model():
#import here, such that our global variables are not overridden!
from . import j6_2ch_128mm, j6_4ch
meta_2ch = j6_2ch_128mm.build_model()
meta_4ch = j6_4ch.build_model()
l_age = nn.layers.InputLayer(data_sizes["sliced:meta:PatientAge"])
l_sex = nn.layers.InputLayer(data_sizes["sliced:meta:PatientSex"])
l_meta_2ch_systole = meta_2ch["meta_outputs"]["systole"]
l_meta_2ch_diastole = meta_2ch["meta_outputs"]["diastole"]
l_meta_4ch_systole = meta_4ch["meta_outputs"]["systole"]
l_meta_4ch_diastole = meta_4ch["meta_outputs"]["diastole"]
l_meta_systole = nn.layers.ConcatLayer(
[l_age, l_sex, l_meta_2ch_systole, l_meta_4ch_systole])
l_meta_diastole = nn.layers.ConcatLayer(
[l_age, l_sex, l_meta_2ch_diastole, l_meta_4ch_diastole])
ldsys1 = nn.layers.DenseLayer(l_meta_systole,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
ldsys1drop = nn.layers.dropout(ldsys1, p=0.5)
ldsys2 = nn.layers.DenseLayer(ldsys1drop,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
ldsys2drop = nn.layers.dropout(ldsys2, p=0.5)
ldsys3 = nn.layers.DenseLayer(ldsys2drop,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
ldsys3drop = nn.layers.dropout(
ldsys3, p=0.5
) # dropout at the output might encourage adjacent neurons to correllate
ldsys3dropnorm = layers.NormalisationLayer(ldsys3drop)
l_systole = layers.CumSumLayer(ldsys3dropnorm)
lddia1 = nn.layers.DenseLayer(l_meta_diastole,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
lddia1drop = nn.layers.dropout(lddia1, p=0.5)
lddia2 = nn.layers.DenseLayer(lddia1drop,
num_units=512,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.rectify)
lddia2drop = nn.layers.dropout(lddia2, p=0.5)
lddia3 = nn.layers.DenseLayer(lddia2drop,
num_units=600,
W=nn.init.Orthogonal("relu"),
b=nn.init.Constant(0.1),
nonlinearity=nn.nonlinearities.softmax)
lddia3drop = nn.layers.dropout(
lddia3, p=0.5
) # dropout at the output might encourage adjacent neurons to correllate
lddia3dropnorm = layers.NormalisationLayer(lddia3drop)
l_diastole = layers.CumSumLayer(lddia3dropnorm)
submodels = [meta_2ch, meta_4ch]
return {
"inputs":
dict(
{
"sliced:meta:PatientAge": l_age,
"sliced:meta:PatientSex": l_sex,
}, **{
k: v
for d in [model["inputs"] for model in submodels]
for k, v in list(d.items())
}),
"outputs": {
"systole": l_systole,
"diastole": l_diastole,
},
"regularizable":
dict({}, **{
k: v
for d in [
model["regularizable"] for model in submodels
if "regularizable" in model
] for k, v in list(d.items())
}),
"pretrained": {
j6_2ch_128mm.__name__: meta_2ch["outputs"],
j6_4ch.__name__: meta_4ch["outputs"],
},
"cutoff_gradients": [] + [
v for d in [
model["meta_outputs"]
for model in submodels if "meta_outputs" in model
] for v in list(d.values())
]
}
