Object Tracking Technology: Trends, Challenges and Applications (Contributions to Environmental Sciences & Innovative Business Technology)

✍ Scribed by Ashish Kumar (editor), Rachna Jain (editor), Ajantha Devi Vairamani (editor), Anand Nayyar (editor)

• Publisher: Springer
• Year: 2023
• Tongue: English
• Leaves: 280
• Edition: 1st ed. 2023
• Category: Library

✦ Synopsis

With the increase in urban population, it became necessary to keep track of the object of interest. In favor of SDGs for sustainable smart city, with the advancement in technology visual tracking extends to track multi-target present in the scene rather estimating location for single target only. In contrast to single object tracking, multi-target introduces one extra step of detection. Tracking multi-target includes detecting and categorizing the target into multiple classes in the first frame and provides each individual target an ID to keep its track in the subsequent frames of a video stream. One category of multi-target algorithms exploits global information to track the target of the detected target. On the other hand, some algorithms consider present and past information of the target to provide efficient tracking solutions. Apart from these, deep leaning-based algorithms provide reliable and accurate solutions. But, these algorithms are computationally slow when applied in real-time. This book presents and summarizes the various visual tracking algorithms and challenges in the domain. The various feature that can be extracted from the target and target saliency prediction is also covered. It explores a comprehensive analysis of the evolution from traditional methods to deep learning methods, from single object tracking to multi-target tracking. In addition, the application of visual tracking and the future of visual tracking can also be introduced to provide the future aspects in the domain to the reader. This book also discusses the advancement in the area with critical performance analysis of each proposed algorithm. This book will be formulated with intent to uncover the challenges and possibilities of efficient and effective tracking of single or multi-object, addressing the various environmental and hardware challenges.
The intended audience includes academicians, engineers, postgraduate students, developers, professionals, military personals, scientists, data analysts, practitioners, and people who are interested in exploring more about tracking.· Another projected audience are the researchers and academicians who identify and develop methodologies, frameworks, tools, and applications through reference citations, literature reviews, quantitative/qualitative results, and discussions.

✦ Table of Contents

Preface
Contents
About the Editors
Single-Object Detection from Video Streaming
1 Introduction
2 Related Work
2.1 Two-Stage-Based Object Detection
2.2 One-Stage-Based Object Detection
3 Materials and Methods
3.1 Materials
3.1.1 Dataset
3.1.2 Data Pre-processing
3.1.3 Data Augmentation
3.1.4 Data Annotation
3.2 Methods
3.2.1 Overview of YOLOv4
4 Applications of YOLOv4
5 Proposed Methodology
6 Experimentation and Implementation
6.1 Experimental Setup
6.2 Training YOLOv4
6.3 Evaluation Measures
7 Results and Performance Analysis
8 Conclusion and Future Work
References
Different Approaches to Background Subtraction and Object Tracking in Video Streams: A Review
1 Introduction
2 Literature Review
2.1 Survey on Frame Rate Conversion Techniques
2.2 Survey on Foreground Extraction Techniques
2.3 Feature Extraction Methods
2.4 Machine Learning Approaches for Pedestrian Detection
2.5 Deep Learning Approaches for Pedestrian Detection
3 Conclusion and Future Scope
References
Auto Alignment of Tanker Loading Arm Utilizing Stereo Vision Video and 3D Euclidean Scene Reconstruction
1 Introduction
2 State of Research in the Field
2.1 Stereo View Geometry
2.2 Geometry of an Epipolar Camera
2.3 Marine Loading Arms
3 Research Methodology
3.1 Disparity Map
3.2 Calibration
3.3 Feature Recognition
3.4 Calibration and Model Fitting
3.5 Distance Calculation
3.6 Reconstruction Error
4 Experimentation and Results
4.1 Extraction of a Specific Target
4.2 Results of Calibration
4.3 Reconstruction Error
4.4 Effects of Errors on 3D Reconstruction
5 Conclusion and Future Scope
References
Visual Object Segmentation Improvement Using Deep Convolutional Neural Networks
1 Introduction
2 Methods for Image Retrieval
2.1 Image Retrieval Techniques
3 Literature Review
4 Data Analysis of Visual Object Segmentation
5 Pre-processing for Anatomical MRI Source Estimation
5.1 Pre-processing for Anatomical Parcellation Labels
6 Source Activity Reconstruction and Encoding Model
7 Nested Cross-Validation and Encoding Linear Model
8 Encoding and Decoding of Pixel Space Control Model
8.1 Improved Cascade Mask R-CNN Model
9 Conclusion and Future Scope
References
Applications of Deep Learning-Based Methods on Surveillance Video Stream by Tracking Various Suspicious Activities
1 Introduction
2 Methods for Detecting Video Anomalies Based on Deep Learning
2.1 Using Patterns as a Global Framework
2.2 Methods Based on Grid Patterns
2.3 Learning Models Based on Representations
2.4 Discriminative Models
2.5 Models of Deep One-Class Categorization
2.6 Models with Deep Hybridization
3 Deep Learning Framework for Detecting Video Anomalies in a Multimodal Semi-supervised Environment
3.1 Overview
3.2 Feature Extraction and Dataset
3.2.1 Extraction of Features
3.2.2 Dataset
3.3 Summary and Discussion
4 Multi-object Tracking
4.1 Overview
4.2 MOT: Procedures, System of Measurement, and Datasets
4.2.1 Overview to MOT Procedures
4.2.2 Metrics
4.3 Deep Learning in MOT
4.3.1 Increased RCNN Speed
4.3.2 SSD
4.3.3 DL in Feature Extraction and Motion Prediction
4.3.4 CNNs as Visual Feature Extractors
4.3.5 Siamese Networks
4.4 Summary and Discussion
5 Video Anomaly Detection and Categorization
5.1 Overview
5.2 RCNN-Based Faster Anomaly Detection
5.3 Anomaly Classification Using DQL
5.4 Summary and Discussion
6 Conclusion and Future Scope
References
Hardware Design Aspects of Visual Tracking System
1 Introduction
2 Challenges in Visual Tracking
2.1 Loss of Information due to Projection of 3D onto 2D
2.2 Noise in Images
2.3 Cluttered Background
2.4 Complex Object Motion
2.5 Partial or Full Occlusions
2.6 Illumination Changes
2.7 Drift Problem
2.8 Requirements of Robust Tracking System
3 Design Flow
4 Common Tracking Algorithms
5 Need for Hardware Design for Visual Tracking
6 Image Processing Algorithms in FPGA Hardware
6.1 Interfacing Camera Sensor to FPGA
6.2 Simple Image Processing Operations
7 Hardware/Software Co-design on FPGA for Visual Tracking
8 Hardware Implementation of Visual Tracking
8.1 Tracker Based on Mean Shift
8.2 Filtering Technique-Based Tracker
8.3 Color-Based Particle Filtering Approach
8.4 Tracker Mechanism Based on Feature Matching
9 Object Tracking Using Multiple Cameras
10 Initial Centering and Fitting: Multiple-Camera Systems (PTZ)
11 Limitations of Hardware Implementation
12 Conclusion and Future Scope
References
Automatic Helmet (Object) Detection and Tracking the Riders Using Kalman Filter Technique
1 Introduction
1.1 Essential Elements of Object Tracking
1.2 Evaluation Criteria for Object Tracking Algorithms
1.3 Key Characteristics of Tracker
1.4 Object Detection Versus Object Tracking
2 Literature Review: Object Detection and Object Tracking
2.1 Feature-Based Tracking
2.2 Model-Based Tracking
2.3 Template Tracking
2.4 Tracking in Compressed Domain
2.5 Tracking with Particle Filter
2.6 Tracking by Background Subtraction
2.7 Tracking Multiple Objects
3 Methodology
4 Methodology Implementation
4.1 Video Segmentation
4.2 Feature Extraction
4.3 Object Detection
4.4 Object Tracking
5 Experimentation Analysis
5.1 Detection of Motorcycles
5.2 Phase I: Identifying Bikers
5.3 Phase II: Identification of Helmet-Less Bikers
5.4 Feature Extraction
5.5 Tracking Using Kalman Filter
6 Conclusion and Future Scope
References
Deep Learning-Based Multi-object Tracking
1 Introduction
2 Object Tracking and Object Recognition
3 Related Works: Deep Learning-Based Tracking Algorithms
4 Overview of Deep Learning-Based MOT
4.1 CNN-Based Algorithm
4.2 Fusion-Based Algorithm
5 Performance Metrics for MOT Algorithm Evaluation
5.1 Description of Performance Metrics
5.2 MOT Performance Indicator
6 Application of Deep Learning in MOT Algorithms
7 Conclusion and Future Scope
References
Multiple Object Tracking of Autonomous Vehicles for Sustainable and Smart Cities
1 Introduction
2 Background
2.1 Sustainable and Smart Cities
2.2 Background to AVs
2.3 AVs with Sustainability
3 Taxonomy of the Multi-object Tracking
3.1 Traditional Multi-object Tracking
3.1.1 Model-Based MOT
3.1.2 Stereo Vision-Based MOT
3.1.3 Grid-Based MOT
3.1.4 Sensor-Fusion Model for MOT
3.2 Deep Learning-Based MOT
3.2.1 Detection-Based vs Detection-Free Multi-object Tracking
3.2.2 Online vs Offline Multi-object Tracking
4 Conclusion and Future Scope
References
Multi-object Detection: A Social Distancing Monitoring System
1 Introduction
1.1 Issues and Challenges for Multi-object Detection
1.2 Need for Computational Analysis
1.3 Applications of Multi-object Detection from Video Streaming Data
1.4 Social Distance Monitoring Issues and Challenges
2 Related Work
3 Application Use Case (Case Study) of Social Distance Monitoring
3.1 Deep Learning
3.2 CNN (Convolutional Neural Network)
3.3 YOLO (You Only Look Once)
3.4 YOLO Algorithm Techniques
3.5 YOLO Architecture
3.6 Formal Model
3.7 Workflow Model
4 Results and Discussion
5 Conclusion and Future Scope
References
Investigating Two-Stage Detection Methods Using Traffic Light Detection Dataset
1 Introduction
2 Overview of DL-Based Models in Object Detection and Tracking
2.1 DL Architectures
2.2 DL Techniques
2.3 Application of DL/CNN in Traffic Light Detection
2.4 Challenges of DL Models in Object Detection and Tracking
3 Traffic Light Detection Using CNN Two-Stage Detection Methods with LARA Pre-trained Dataset
3.1 Data Collection
3.2 Preprocessing of TLD Dataset
3.3 Selection of Features
3.4 The Generic Object Detector
3.5 Evaluation Methods
3.6 Experimentation Setup
4 Results and Discussion
5 Conclusion and Future Scope
References