Deep Learning for Medical Applications with Unique Data

By Deepak Gupta, Utku Kose and Valentina Emilia Balas

Deep Learning for Medical Applications with Unique Data informs readers about the most recent deep learning-based medical applications in which only unique data gathered in real cases are used. The book provides examples of how deep learning can be used in different problem areas and frameworks in both clinical and research settings, including medical image analysis, medical image registration, time series analysis, medical data synthesis, drug discovery, and pre-processing operations. The volume discusses not only positive findings, but also negative ones obtained by deep learning techniques, including the use of newly developed deep learning techniques rarely reported in the existing literature. The book excludes research works with ready data sets and includes only unique data use to better understand the state of deep learning in real-world cases, along with the feedback and user experiences from physicians and medical staff for applied deep learning-based solutions. Other applications presented in the book include hybrid solutions with deep learning support, disease diagnosis with deep learning focusing on rare diseases and cancer, patient care and treatment, genomics research, as well as research on robotics and autonomous systems.

• Introduces deep learning, demonstrating concepts for a wide variety of medical applications using unique data, excluding research with ready datasets
• Encompasses a wide variety of biomedical applications, including unsupervised learning, natural language processing, pattern recognition, image and video processing and disease diagnosis
• Provides a robust set of methods that will help readers appropriately and judiciously use the most suitable deep learning techniques for their applications

• Language: English
• Publisher: Academic Press
• Release date: Feb 15, 2022
• ISBN: 9780128241462

Book preview

Preface

Developments in the field of medicine have been always associated with the intense use of technology. As a result of using knowledge to produce effective technological tools, it has been possible for humanity to deal with critical problems in medicine. The effective use of computers and communication technology has resulted in new approaches to critical problems such as medical diagnosis, medical treatment, drug discovery, and precision medicine. Among the different technological tools, artificial intelligence is a revolutionary invention because it employs adaptive, advanced algorithms able to respond to changing data and even learn from them to produce descriptive or predictive outcomes. Deep learning is the most valuable subfield of artificial intelligence. It is widely used to deal with complicated medical problems including different types of data. The use of machine and deep learning models is an essential way to solve medical problems, but another important requirement in solving target problems is employing the most appropriate dataset.

Whether or not they are in the context of medical problems, it is necessary to have datasets to run artificial intelligence–based applications. Datasets are the digital transformation of collected knowledge so that an endless loop between technological solutions and new knowledge or information is ensured. Data sets for traditional machine learning models in the 20th century required effective preprocessing and the careful collection of associated data for target problems. This has been transformed into the accurate use of a large and complex amount of data, owing to deep learning models. It is remarkable that deep learning models have been more successful than traditional machine learning models when it comes to dealing with complex problems with more data. However, even deep learning needs some preprocessing and careful gathering of data to produce improved outcomes for different types of problems. From raw to signal-based data and from mixed data to image data, deep learning models in medical applications enable researchers to work on datasets before passing to exact applications. Thus, datasets and the changing nature of problems according to the chosen datasets are popular focuses of research in the intersection of artificial intelligence and the medical field.

This edited book, Deep Learning for Medical Applications with Unique Data, gathers research work, including the effective use of deep learning models, to solve critical medical problems through datasets. In detail, the chapters were chosen according to their unique value for target problems and the datasets used. To understand the method of artificial intelligence and deep learning in different medical problems, it is important to learn from research and the results that are obtained. Thus, the chapters included in this book specifically focus on problems such as brain tumor detection, cell detection, COVID-19 diagnosis, early heart disease prediction, and the diagnosis of glaucoma. In addition, the book was improved by enabling the authors to provide a deep review of the associated literature so that readers are able to obtain information about deep learning and medical dataset synergy.

We believe that readers, including students, scientists from different fields, and experts and professionals from public and private sectors, will benefit greatly from the contents of this book. As the editors, we would like to thank all of the authors and the Elsevier team, who showed great effort in developing this book. We are also looking forward to receiving contributive feedback, ideas about the volume, and alternative research topics that readers believe that we should edit in future projects. Most sincerely.

Editors

Assist. Prof. Dr. Deepak Gupta

Department of Computer Science and Engineering,

Maharaja Agrasen Institute of Technology (MAIT),

New Delhi, India

https://sites.google.com/view/drdeepakgupta/home

[e-mail]

Assoc. Prof. Dr. Utku Kose

Department of Computer Engineering,

Süleyman Demirel University, Isparta, Turkey

http://www.utkukose.com/

[e-mail]

Assist. Prof. Dr. Ashish Khanna

Department of Computer Science and Engineering,

Maharaja Agrasen Institute of Technology (MAIT),

New Delhi, India

[e-mail]

Prof. Dr. Valentina Emilia Balas

Department of Automatics and Applied Software,

Faculty of Engineering, Aurel Vlaicu University of Arad,

Arad, Romania

https://www.drbalas.ro/

[e-mail]

1: A deep learning approach for the prediction of heart attacks based on data analysis

C.V. Aravinda ¹ , Meng Lin ² , K.R. Udaya Kumar Reddy ³ , and G. Amar Prabhu ⁴       ¹ N.M.A.M. Institute of Technology Nitte, Karkala, India      ² Ritsumeikan University, Kusatsu, Shiga, Japan      ³Dayananda Sagar College of Engineering, Bengaluru, India      ⁴ Komatsu Kaihatsu Company, Kariya, Aichi, Japan

Abstract

Coronary illness has a wide range of conditions that influences the heart. It is one of the most complex disorders to predict because of the number of components in the body that could prompt it. Distinguishing and anticipating it are challenging for specialists and analysts. If we analyze deaths caused by those that were avoidable those from other reasons in India, the third most prevalent cause of unnecessary death is due to heart attack. Across the globe, number of these types of death may increase to more than 23.6 million by 2030. Around 80% of deaths caused by heart attack occur mainly to younger people. This chapter supports medical specialists in detecting and predicting heart disease by attaining precision levels, as well as in prescribing effective medicine according to the disease findings. Given sensor data, deep learning algorithms are applied along with neural network, random forest, and decision tree classifiers to analyze patients' data to predict heart disease. The experiment shows that the prediction of heart disease has promising results with about 90% accuracy.

Keywords

Artificial neural network; Decision tree; Naive Bayes; Neural network; Random forest

1. Introduction

2. Literature survey

3. Materials and method

3.1 Cohort study

4. Training models

4.1 Artificial neural network

4.2 K-nearest neighbor classifier

4.3 Naive Bayes classifier

4.4 Decision tree classifier

4.5 Random forest