This book provides state-of-the-art coverage of deep learning applications in image analysis. The book demonstrates various deep learning algorithms that can offer practical solutions for various image-related problems; also how these algorithms are used by scientists and scholars in industry and academia. This includes autoencoder and deep convolutional generative adversarial network in improving classification performance of Bangla handwritten characters, dealing with deep learning-based approaches using feature selection methods for automatic diagnosis of covid-19 disease from x-ray images, imbalance image data sets of classification, image captioning using deep transfer learning, developing a vehicle over speed detection system, creating an intelligent system for video-based proximity analysis, building a melanoma cancer detection system using deep learning, plant diseases classification using AlexNet, dealing with hyperspectral images using deep learning, chest x-ray image classification of pneumonia disease using efficient net and inceptionv3.
The book also addresses the difficulty of implementing deep learning in terms of computation time and the complexity of reasoning and modelling different types of data where information is currently encoded. Each chapter has the application of various new or existing deep learning models such as Deep Neural Network (DNN) and Deep Convolutional Neural Networks (DCNN). The detailed utilization of deep learning packages that are available in MATLAB, Python and R programming environments have also been discussed, therefore, the readers will get to know about the practical implementation of deep learning as well. The content of this book is presented in a simple and lucid style for professionals, nonprofessionals, scientists, and students interested in the research area of deep learning applications in image analysis.
Author(s): Sanjiban Sekhar Roy, Ching-Hsien Hsu, Venkateshwara Kagita
Series: Studies in Big Data, 129
Publisher: Springer
Year: 2023
Language: English
Pages: 217
City: Singapore
Preface
Contents
About the Editors
Autoencoder and Deep Convolutional Generative Adversarial Network in Improving the Performance of Bangla Handwritten Character Recognition
1 Introduction
2 Related Work
3 Materials and Methods
3.1 Dataset
3.2 Outlier Detection
3.3 Generative Adversarial Network
3.4 Classification
3.5 Train, Test, and Validation
4 Results
4.1 Result Analysis
4.2 Overfitting Handling
4.3 Comparison with State-of-the-Art
5 Discussion
6 Conclusion
References
Deep Learning-Based Approaches Using Feature Selection Methods for Automatic Diagnosis of COVID-19 Disease from X-Ray Images
1 Introduction
2 Material and Method
2.1 Methodology
2.2 Dataset
3 Performance Measurement Metrics
4 Experimental Studies
5 Discussion
6 Results
References
Image Captioning Using Deep Transfer Learning
1 Introduction
2 Related Studies
3 Methodology
3.1 Dataset
3.2 Inception Model for Images
4 Result
4.1 Sample Output
5 Conclusion
References
Vehicle Over Speed Detection System
1 Introduction
2 Proposed Model
2.1 Vehicle Detection and License Plate Extraction System
2.2 Travel Time Estimating and Over-Speed Detection System
3 Result
3.1 Vehicle Detection and License Plate Localization and Text Extraction
3.2 New Curve Detection Algorithm
3.3 Curve Aware Travel Time Estimation
3.4 Vehicle Over Speed Detection System
3.5 Discussion
4 Conclusion
References
An Intelligent System for Video-Based Proximity Analysis
1 Introduction
2 Overview of the Framework
3 Construction of Proximity Networks
3.1 People Detection and Accuracy Evaluation
3.2 Finding Coordinates of Each Person
3.3 Extracting Walking Trajectories
3.4 Finding People That Appear in Close Proximity
3.5 Face Analysis and Recognition
4 Experiments
4.1 Combined Dataset of Neural Network Training
4.2 Training a Neural Network Model
4.3 Frame Processing
4.4 Contact Network Graph
5 Further Interpretation and Outlook Towards Adaptation to the Post-pandemic Society Goals
6 Conclusion and Outlook
References
Deep Learning-Based Conjunctival Melanoma Detection Using Ocular Surface Images
1 Introduction
2 Methodology
2.1 Data Collection
2.2 Data Augmentation
2.3 Convolutional Neural Network (CNN) Based Classification Models
2.4 Visualization Methods
2.5 Experimental Setup
2.6 Evaluation Metrics
3 Results
3.1 Binary Classification
3.2 Multi-class Classification
3.3 Comparative Analysis with Existing Literature
3.4 Visualization Using Grad-CAM
4 Conclusion
References
Plant Diseases Classification Using Neural Network: AlexNet
1 Introduction
2 Machine Learning and Deep Learning
2.1 History
2.2 Machine Learning Basics
2.3 Convolution Neural Network
2.4 Various Deep Learning Libraries
3 Experimental Work and Results
3.1 Dataset
3.2 Data Pre-processing
3.3 Architecture
3.4 Results
4 Conclusion
References
Hyperspectral Images: A Succinct Analytical Deep Learning Study
1 Introduction
2 Related Researches
3 Hyperspectral Images
4 Deep Learning and CNN
4.1 HI Based Deep Feature Selection
4.2 HI Based Optimization
5 3D-Convolutional Neural Network Based HI Classification on Sentinel-2 Satellite Data of Sundarban Mangrove Regions
5.1 Dataset Description
5.2 Experimental Setup and Hyper-Parameters
6 A Novel Deep Learning Hybrid-MSSN Architecture for Hyperspectral Image Classification
6.1 Architecture of Hybrid-MSSN Model
6.2 Dataset Description
6.3 Experimental Setup and Result Analysis
7 Conclusions and Future Scope
References
Chest X-Ray Image Classification of Pneumonia Disease Using EfficientNet and InceptionV3
1 Introduction
2 Literature Survey
3 Dataset
4 Data Pre-processing
5 Proposed Model
5.1 EfficientNet
5.2 InceptionV3
6 Experimental Outcome and Analysis
6.1 InceptionV3
6.2 EfficientNet
7 Discussion
8 Conclusion
References
Detection of Cancer Using Deep Learning Techniques
1 Introduction
2 Deep Learning
2.1 Basics of Deep Learning
2.2 Cancer Diagnosis with DL
2.3 Deep Neural Network Characteristics
3 Architectures of Deep Learning Neural Networks
3.1 Deep Unsupervised Learning
3.2 Deep Supervised Learning
3.3 Deep Semi-supervised Learning
4 Types of Deep Learning Architectures for Cancer Detection
4.1 Convolutional Neural Networks (CNN)
4.2 Multi-scale Convolution Neural Network
4.3 LeNet-5
4.4 AlexNet
4.5 ZFNet
4.6 GoogleNet
4.7 VGGNet
4.8 ResNet
4.9 Fully Convolutional Networks (FCNs)
4.10 U-Net
4.11 Recurrent Neural Networks
4.12 Autoencoders
4.13 Deep Belief Networks
5 Steps for Diagnosis of Cancer by Medical Imaging
5.1 Cleaning and Pre-processing
5.2 Image Segmentation
5.3 Post Processing
6 Diagnosis of Different Types of Cancers Using DL
7 Conclusions
References
