Bases: XAImethod_Base
XAI Method for Guided Backpropagation.
Normally, this class is used internally in the aucmedi.xai.decoder.xai_decoder in the AUCMEDI XAI module.
Reference - Implementation #1
Author: Hoa Nguyen
GitHub Profile: https://nguyenhoa93.github.io/
Date: Jul 29, 2020
https://stackoverflow.com/questions/55924331/how-to-apply-guided-backprop-in-tensorflow-2-0
Reference - Implementation #2
Author: Huynh Ngoc Anh
GitHub Profile: https://github.com/experiencor
Date: Jun 23, 2017
https://github.com/experiencor/deep-viz-keras/
Reference - Implementation #3
Author: Tim
Date: Jan 25, 2019
https://stackoverflow.com/questions/54366935/make-a-deep-copy-of-a-keras-model-in-python
Reference - Publication
Jost Tobias Springenberg, Alexey Dosovitskiy, Thomas Brox, Martin Riedmiller. 21 Dec 2014.
Striving for Simplicity: The All Convolutional Net.
https://arxiv.org/abs/1412.6806
This class provides functionality for running the compute_heatmap function,
which computes a Guided Backpropagation for an image with a model.
Source code in aucmedi/xai/methods/guided_backprop.py
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129 | class GuidedBackpropagation(XAImethod_Base):
""" XAI Method for Guided Backpropagation.
Normally, this class is used internally in the [aucmedi.xai.decoder.xai_decoder][] in the AUCMEDI XAI module.
??? abstract "Reference - Implementation #1"
Author: Hoa Nguyen <br>
GitHub Profile: [https://nguyenhoa93.github.io/](https://nguyenhoa93.github.io/) <br>
Date: Jul 29, 2020 <br>
[https://stackoverflow.com/questions/55924331/how-to-apply-guided-backprop-in-tensorflow-2-0](https://stackoverflow.com/questions/55924331/how-to-apply-guided-backprop-in-tensorflow-2-0) <br>
??? abstract "Reference - Implementation #2"
Author: Huynh Ngoc Anh <br>
GitHub Profile: [https://github.com/experiencor](https://github.com/experiencor) <br>
Date: Jun 23, 2017 <br>
[https://github.com/experiencor/deep-viz-keras/](https://github.com/experiencor/deep-viz-keras/) <br>
??? abstract "Reference - Implementation #3"
Author: Tim <br>
Date: Jan 25, 2019 <br>
[https://stackoverflow.com/questions/54366935/make-a-deep-copy-of-a-keras-model-in-python](https://stackoverflow.com/questions/54366935/make-a-deep-copy-of-a-keras-model-in-python) <br>
??? abstract "Reference - Publication"
Jost Tobias Springenberg, Alexey Dosovitskiy, Thomas Brox, Martin Riedmiller. 21 Dec 2014.
Striving for Simplicity: The All Convolutional Net.
<br>
[https://arxiv.org/abs/1412.6806](https://arxiv.org/abs/1412.6806)
This class provides functionality for running the compute_heatmap function,
which computes a Guided Backpropagation for an image with a model.
"""
def __init__(self, model, layerName=None):
""" Initialization function for creating Guided Backpropagation as XAI Method object.
Args:
model (keras.model): Keras model object.
layerName (str): Not required in Guided Backpropagation, but defined by Abstract Base Class.
"""
# Create a deep copy of the model
model_copy = tf.keras.models.clone_model(model)
model_copy.build(model.input.shape)
model_copy.compile(optimizer=model.optimizer, loss=model.loss)
model_copy.set_weights(model.get_weights())
# Define custom Relu activation function
@tf.custom_gradient
def guidedRelu(x):
def grad(dy):
return tf.cast(dy>0, "float32") * tf.cast(x>0, "float32") * dy
return tf.nn.relu(x), grad
# Replace Relu activation layers with custom Relu activation layer
layer_dict = [layer for layer in model_copy.layers if hasattr(layer, "activation")]
for layer in layer_dict:
if layer.activation == tf.keras.activations.relu:
layer.activation = guidedRelu
# Cache class parameters
self.model = model_copy
#---------------------------------------------#
# Heatmap Computation #
#---------------------------------------------#
def compute_heatmap(self, image, class_index, eps=1e-8):
""" Core function for computing the Guided Backpropagation for a provided image and for specific classification outcome.
???+ attention
Be aware that the image has to be provided in batch format.
Args:
image (numpy.ndarray): Image matrix encoded as NumPy Array (provided as one-element batch).
class_index (int): Classification index for which the heatmap should be computed.
eps (float): Epsilon for rounding.
The returned heatmap is encoded within a range of [0,1]
???+ attention
The shape of the returned heatmap is 2D -> batch and channel axis will be removed.
Returns:
heatmap (numpy.ndarray): Computed Guided Backpropagation for provided image.
"""
# Compute gradient for desierd class index
with tf.GradientTape() as tape:
inputs = tf.cast(image, tf.float32)
tape.watch(inputs)
preds = self.model(inputs)
loss = preds[:, class_index]
gradient = tape.gradient(loss, inputs)
# Obtain maximum gradient based on feature map of last conv layer
gradient = tf.reduce_max(gradient, axis=-1)
# Convert to NumPy & Remove batch axis
heatmap = gradient.numpy()[0,:,:]
# Intensity normalization to [0,1]
numer = heatmap - np.min(heatmap)
denom = (heatmap.max() - heatmap.min()) + eps
heatmap = numer / denom
# Return the resulting heatmap
return heatmap
|
__init__(model, layerName=None)
Initialization function for creating Guided Backpropagation as XAI Method object.
Parameters:
| Name |
Type |
Description |
Default |
model |
keras.model
|
Keras model object. |
required
|
layerName |
str
|
Not required in Guided Backpropagation, but defined by Abstract Base Class. |
None
|
Source code in aucmedi/xai/methods/guided_backprop.py
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87 | def __init__(self, model, layerName=None):
""" Initialization function for creating Guided Backpropagation as XAI Method object.
Args:
model (keras.model): Keras model object.
layerName (str): Not required in Guided Backpropagation, but defined by Abstract Base Class.
"""
# Create a deep copy of the model
model_copy = tf.keras.models.clone_model(model)
model_copy.build(model.input.shape)
model_copy.compile(optimizer=model.optimizer, loss=model.loss)
model_copy.set_weights(model.get_weights())
# Define custom Relu activation function
@tf.custom_gradient
def guidedRelu(x):
def grad(dy):
return tf.cast(dy>0, "float32") * tf.cast(x>0, "float32") * dy
return tf.nn.relu(x), grad
# Replace Relu activation layers with custom Relu activation layer
layer_dict = [layer for layer in model_copy.layers if hasattr(layer, "activation")]
for layer in layer_dict:
if layer.activation == tf.keras.activations.relu:
layer.activation = guidedRelu
# Cache class parameters
self.model = model_copy
|
compute_heatmap(image, class_index, eps=1e-08)
Core function for computing the Guided Backpropagation for a provided image and for specific classification outcome.
Attention
Be aware that the image has to be provided in batch format.
Parameters:
| Name |
Type |
Description |
Default |
image |
numpy.ndarray
|
Image matrix encoded as NumPy Array (provided as one-element batch). |
required
|
class_index |
int
|
Classification index for which the heatmap should be computed. |
required
|
eps |
float
|
Epsilon for rounding. |
1e-08
|
The returned heatmap is encoded within a range of [0,1]
Attention
The shape of the returned heatmap is 2D -> batch and channel axis will be removed.
Returns:
| Name | Type |
Description |
heatmap |
numpy.ndarray
|
Computed Guided Backpropagation for provided image. |
Source code in aucmedi/xai/methods/guided_backprop.py
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129 | def compute_heatmap(self, image, class_index, eps=1e-8):
""" Core function for computing the Guided Backpropagation for a provided image and for specific classification outcome.
???+ attention
Be aware that the image has to be provided in batch format.
Args:
image (numpy.ndarray): Image matrix encoded as NumPy Array (provided as one-element batch).
class_index (int): Classification index for which the heatmap should be computed.
eps (float): Epsilon for rounding.
The returned heatmap is encoded within a range of [0,1]
???+ attention
The shape of the returned heatmap is 2D -> batch and channel axis will be removed.
Returns:
heatmap (numpy.ndarray): Computed Guided Backpropagation for provided image.
"""
# Compute gradient for desierd class index
with tf.GradientTape() as tape:
inputs = tf.cast(image, tf.float32)
tape.watch(inputs)
preds = self.model(inputs)
loss = preds[:, class_index]
gradient = tape.gradient(loss, inputs)
# Obtain maximum gradient based on feature map of last conv layer
gradient = tf.reduce_max(gradient, axis=-1)
# Convert to NumPy & Remove batch axis
heatmap = gradient.numpy()[0,:,:]
# Intensity normalization to [0,1]
numer = heatmap - np.min(heatmap)
denom = (heatmap.max() - heatmap.min()) + eps
heatmap = numer / denom
# Return the resulting heatmap
return heatmap
|