Читать книгу Machine Learning for Healthcare Applications - Группа авторов - Страница 4

1 Chapter 1Figure 1.1 Categorization of healthcare fraud.

2 Chapter 2Figure 2.1 Architecture of the model.Figure 2.2 Screenshots of the web application.Figure 2.3 Accuracy: Model-I vs Model-II.Figure 2.4 Precision: Model-I vs Model-II.Figure 2.5 Recall: Model-I vs Model-II.Figure 2.6 Recall: Model-I vs Model-II.

3 Chapter 3Figure 3.1 Brain map structure and Equipment used.Figure 3.2 Workflow diagram.Figure 3.3 DWT schematic.Figure 3.4 Images used for visual evaluation.Figure 3.5 Sample of EEG signal for a product with corresponding Brain map and c...Figure 3.6 Accuracy for all users (compiled).Figure 3.7 Individual result of each algorithm.Figure 3.8 Result of 25-users with different algorithms.Figure 3.9 Result of 25-users compared with different algorithms.Figure 3.10 Approximate brain EEG map for dislike state.Figure 3.11 Approximate brain EEG map for like state.

4 Chapter 4Figure 4.1 Classification of clinical DSS.Figure 4.2 Architecture of CDSS [29].Figure 4.3 Inference using decision tree for the proposed system.Figure 4.4 (a) First level UI of the system in ES-Builder.Figure 4.4 (b) Second level UI of the system in ES-Builder.Figure 4.4 (c) Third level UI of the system in ES-Builder.Figure 4.4 (d) Fourth level UI of the system in ES-Builder.Figure 4.4 (e) Fifh level UI of the system in ES-Builder.Figure 4.4 (f) Sixth level UI of the system in ES-Builder.Figure 4.4 (g) Conclusion level UI of the system in ES-Builder.

5 Chapter 5Figure 5.1 Graphs: (a) Euclidean graph, (b) Non-euclidean graph.Figure 5.2 Representation of DSN.Figure 5.3 Training steps of the model.Figure 5.4 Training Performance: (a) Loss, (b) Accuracy.Figure 5.5 Performance comparison: (a) Accuracy, (b) Precision, (c) Recall, (d) ...

6 Chapter 7Figure 7.1 Sample association between URI, RDF and SPARQL.Figure 7.2 SKCE Multi agent system flowchart.Figure 7.3 Semantic translation framework for healthcare instance data.Figure 7.4 Sample data dictionary with meta classes, concepts and concept values...Figure 7.5 Concept level mappings between different data dictionary elements.Figure 7.6 Sample RDF model of concept level mapping between different data mode...

7 Chapter 8Figure 8.1 The tDCS montage was 7x5 mm electrodes centred over F3 (connector at ...Figure 8.2 (a) Spectral Decomposition of EEG of Subject 8 which shows bad channe...Figure 8.3 Legendre Spectrum of Subject 8 for Anodal Session.Figure 8.4 ‘Stimuli’ is coded as ‘0’ for subjects given at their first visit to ...Figure 8.5 Relationships between pair of variables, in the form of a 6 × 6 matri...Figure 8.6 (a) While “SESSION” is coded from 1 to 6 for Anodal-Pre, Anodal-DCS, ...Figure 8.7 (a) “Gender” is coded in such a way that ‘0’ denotes a Male and ‘1’ d...

8 Chapter 9Figure 9.1 Human wrist characterization through the microwave setup.Figure 9.2 Transfer characteristics through the simulated wrist for standard siz...Figure 9.3 Simulated transfer characteristic with healthy bone by varying the bo...Figure 9.4 Simulated transfer characteristic with osteopenia bone by varying the...Figure 9.5 Simulated transfer characteristic with osteoporotic bone 1 by varying...Figure 9.6 Simulated transfer characteristic with osteoporotic bone 2 by varying...Figure 9.7 Simulated transfer characteristic with osteoporotic bone 3 by varying...Figure 9.8 Simulated transfer characteristic with osteoporotic bone 4 by varying...Figure 9.9 Confusion matrix for KNN.Figure 9.10 Confusion matrix for decision tree.Figure 9.11 Confusion matrix for random forest.Figure 9.12 Graphical representation of the classification report.

9 Chapter 10Figure 10.1 Industry landscape of AI in healthcare (Courtesy: Emily Kuo [2]).Figure 10.2 Tomra—Tomato sorting and processing machines (Courtesy: Tomra [9]).Figure 10.3 Kankan’s Machine system (Courtesy: KanKan AI [11]).Figure 10.4 Plant disease detection (Courtesy: Bitrefine [12]).Figure 10.5 Lameness of domestic cattle (Courtesy: Shearer et al. [13]).Figure 10.6 Processing steps in sentiment analysis.Figure 10.7 Bi-direction LSTM model for text sequence classification.Figure 10.8 Word embedding representation in vector space (Courtesy: David Rozad...Figure 10.9 BERT input embeddings (Courtesy: Cheney [25]).Figure 10.10 Fine-tuning of pre-trained BERT models.Figure 10.11 BERT layered model with classifier (Courtesy: Chris McCormick and N...

10 Chapter 11Figure 11.1 A snapshot of HAM10000 dataset.Figure 11.2 A high-level view of our classification model.Figure 11.3 Training and validation loss MobileNet.Figure 11.4 Training and validation loss ResNet50.Figure 11.5 Training and validation categorical accuracy MobileNet.Figure 11.6 Training and validation categorical accuracy ResNet50.Figure 11.7 Training and validation top2 accuracy MobileNet.Figure 11.8 Training and validation top2 accuracy ResNet50.Figure 11.9 Training and validation top3 accuracy MobileNet.Figure 11.10 Training and validation top3 accuracy ResNet50.Figure 11.11 Confusion matrix MobileNet.Figure 11.12 Confusion matrix ResNet50.Figure 11.13 Classification reports MobileNet.Figure 11.14 Classification reports ResNet50.Figure 11.15 Last epoch results MobileNet.Figure 11.16 Last epoch results ResNet50.Figure 11.17 Best epoch results MobileNet.Figure 11.18 Best epoch results ResNet50.

11 Chapter 12Figure 12.1 Globally vulnerable areas affected by malaria.Figure 12.2 Block diagram for proposed work.Figure 12.3 Dataset sample images.Figure 12.4 Classes distribution in training set.Figure 12.5 Classes distribution in validation set.Figure 12.6 CNN architecture.Figure 12.7 Accuracy curve.Figure 12.8 Loss curve.Figure 12.9 Normalized confusion matrix.

12 Chapter 13Figure 13.1 Sample input images of lung diseases.Figure 13.2 Histogram representation of the dataset.Figure 13.3 Output of augmentation process.Figure 13.4 Lung disease prediction model.Figure 13.5 The proposed layer construction.Figure 13.6 Calculation of model parameters.Figure 13.7 Training history of first round.Figure 13.8 ROC curve of first round.Figure 13.9 Training history of final round.Figure 13.10 ROC curve of second round.Figure 13.11 (a), (b), (c), (d), (e), (f), (g), (h) Prediction results—Lung dise...

13 Chapter 14Figure 14.1 Various methods for detecting Leukemia.Figure 14.2 Normal blood and Leukemia infected blood.Figure 14.3 Basic Block diagram of proposed methodology.Figure 14.4 Flowchart of implemented modules.

14 Chapter 15Figure 15.1 Complete layout of the network systems using IoT.Figure 15.2 Network layer of IoT systems.Figure 15.3 List of acronyms and their definitions.Figure 15.4 Smart hospital layout.Figure 15.5 Objectives of smart hospitals.Figure 15.6 Assets of smart hospitals.

15 Chapter 16Figure 16.1 Block diagram of the proposed model.Figure 16.2 Classification using hyperplane.Figure 16.3 Work flow of the health monitoring system.Figure 16.4 Performance analysis using sensitivity.Figure 16.5 Performance analysis using specificity.Figure 16.6 Performance analysis based on accuracy (%).Figure 16.7 Comparative of performance analysis.Figure 16.8 Overview of architecture and interfaces of a system.Figure 16.9 Use case diagram of system design.Figure 16.10 sequence diagram of system design.

16 Chapter 17Figure 17.1 Ontology development and information retrieval process.Figure 17.2 Snippet of concept hierarchy.Figure 17.3 Visualization of concept hierarchy (sample case).Figure 17.4 Information retrieval from knowledgebase.

17 Chapter 18Figure 18.1 Data flow diagram of COVID-19 sentence classification.

18 Chapter 19Figure 19.1 A graph of two class problem with linear separable hyper-plane [21,2...Figure 19.2 Flowchart of the SVM model [23].Figure 19.3 Flowchart of decision tree model [6,24].Figure 19.4 Flowchart of k-means clustering algorithm [6,25].Figure 19.5 Flowchart of levenberg maquardt (LM) training algorithm [6,28].Figure 19.6 Accuracy comparisonFigure 19.7 Training state of levenberg maquardt (LM) method.Figure 19.8 Risk wise classification of other well-known countries.

19 Chapter 20Figure 20.1 Lung CT scan image.Figure 20.2 Crossover.Figure 20.3 Mutation.Figure 20.4 Proposed diagnostic system architecture.Figure 20.5 Original lung CT image.Figure 20.6 Segmentation—Proposed work.Figure 20.7 Lung CT image—segmented.Figure 20.8 ROI segmentation.

20 Chapter 21Figure 21.1 Stages for lung sound prediction.Figure 21.2 The circuit structure of the decomposition.Figure 21.3 Modified random forest architecture for LSS.Figure 21.4 Various performance measures with db4 and MFCC feature extraction of...Figure 21.5 Performance of the Modified-RF, AdaBoost, GB classification algorith...

21 Chapter 22Figure 22.1 Sample sputum smear TB images and its ground truth (taken from Refs....Figure 22.2 Sample bacilli patches used for training and testing. Bacilli images...Figure 22.3 Architecture of the benchmark model (left) and the proposed model (r...Figure 22.4 Training loss and accuracy (mean of 10 experiments) of the proposed ...

22 Chapter 23Figure 23.1 Four patches manually cropped from the image.Figure 23.2 Sixteen sample patches from each class.Figure 23.3 Workflow of the case study for the diagnosis of laryngeal cancer.

23 Chapter 24Figure 24.1 An example of SVM.Figure 24.2 Computer aided diagnosis diagram for diabetic retinopathy detection ...Figure 24.3 Some fundus images.Figure 24.4 Some images of hard exudates and hemorrhages.Figure 24.5 Some images of soft exudates and res small dots.Figure 24.6 Some images of the left eye.Figure 24.7 Some images of the right eye.Figure 24.8 Comparison of accuracy among different classifiers.