Page 304 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

Index 295

segmentation architectures, virtual assistants, 82 tools, 63e68
36e37 visual recognition, 82 two-dimensional tensor,
VGGNet, 163e164, 164f architectures, 284e285 64e65
Convolutional neural networks artificial neural networks, 193f type, 64
(CNNs), 156 basic architecture, 27f Deep neural network
architecture, 5f branches, 196e197 artificial neural network, 283
artificial intelligence (AI), 3f challenges, 194e195 backpropagation algorithm,
coronavirus (COVID-19), 2 Convolutional Architecture for 271
fire module, 9e11 Fast Feature Embedding computed tomography (CT)
security, 11e12 (CAFFE), 67 images, 273
SqueezeNet architecture, 4f features, 67 computer vision, 274e283
training and prediction theano, 68 deep learning, 283e285
scheme, 4f torch tool, 68 digital image, 274e283, 276f
types, 2t diagnostics, 198e208 enhancement, 279
Convolution layerekernel, 74, challenges, 203e207 image processing steps,
74f future, 207e208 277e280, 278f
hominoid barricades, 206 medical image formats,
Data set, 158, 158fe159f motivation, 200e202 275e277
Deep belief networks (DBNs), retroactive versus noncontextual
69e71, 69f forthcoming trainings, segmentation, 279
architecture, 70f 203e205 object recognition,
fine-tuning stage, 255 solutions, 203e207 279e280
greedy layer-wise trouble associating segmentation, 279
unsupervised training dissimilar algorithms, 206 GoogLeNet, 272
principle, 69e70 vulnerability, 207 labeled data set, 281
pretraining stage, 254 early detection of diseases, machine learning, 280, 280f
restricted Boltzmann machine 208e214 reinforcement learning,
(RBM), 69 Alzheimer’s disease, 282e283
visible joint distribution, 69f 210e211 supervised learning, 281
working, 71 attention deficit unlabeled data set, 281
working of, 71 hyperactivity disorder unsupervised learning, 282
Deep Boltzmann Machine (ADHD), 213e214 Depression
(DBM), 81e82 autism spectrum disorder, continuous bag of words
Deep learning (DL), 25 212e213 (CBOW), 139e141, 140f
algorithms, 69e82 rheumatic diseases, convolutional neural network,
Amazon Web Service (AWS), 209e210 142e144, 142fe143f
66e67 further advancements, 214 bidirectional long short-
applications, 195f Keras, 65e67 term memory, 147e148,
automatic handwriting backend, 65e66 148f
generation, 83 TensorFlow vs., 66t, 63 long short-term memory,
automatic machine machine learning (ML), 146e148, 146f
translation, 83, 194 192 multichannel CNN model,
entertainment, 82 one-dimensional tensor, 64 144
fraud news detection, 193 rank, 63 nonstatic-CNN model, 144
imageelanguage rectified linear unit (ReLU), 62 random-CNN model, 144
translations, 83 shape, 64 recurrent neural network,
natural language processing sigmoid function, 62f 144e146, 144fe145f
(NLP), 194 tensor data structure, 63 static-CNN model, 144
self-driving cars, 193 TensorFlow, 63e65 variants, 144