Deep Learning Methods For Image Object Detection And Recognition

This book provides a comprehensive introduction on applying deep learning to the visual recognition of objects, bridging the gap between theoretical algorithms and high-impact practical implementations. Within these pages, readers will find a detailed exploration of advanced techniques for image restoration—including UNet-based defogging, feature fusion GANs, and ESRGAN super-resolution—alongside high-performance detection models. Specialized applications such as underwater crack segmentation using transfer learning, marine biological detection, and real-time analysis through YOLOv4, RetinaNet, and LSTM-integrated networks are also covered to provide researchers and engineers with a technical blueprint for solving diverse real-world computer vision challenges. Additionally, this work is also suitable as a textbook for advanced undergraduate and graduate students majoring in Artificial Intelligence, Intelligent Science and Technology, Computer Science, and Automation.

Pengfei Shi, PhD, is an Associate Professor and Master Supervisor at Hohai University. He is selected into the "Dayu Scholar" program of Hohai University, a member of the CAA Networked Intelligence Special Committee, a member of CCF, and a member of IEEE. He has been long engaged in research in the fields of artificial intelligence and robotics, smart power systems, smart water conservancy, edge computing, machine vision, software and hardware development, etc. He has presided over 1 National Natural Science Foundation project, 1 sub-project of the National Key Research and Development Program, 1 Industry-University-Research Collaborative Education Project of the Ministry of Education, 2 Natural Science Foundation projects of Jiangsu Province, 2 Basic Research Programs of Changzhou City, and multiple entrusted scientific and technological projects from enterprises and institutions. He has published more than 100 papers, among which more than 50 are retrieved by SCI/EI, including many top journal papers and top conference papers in the fields of artificial intelligence and machine vision, such as IEEE Trans. on Instrumentation and Measurement and ICCV. He has applied for more than 30 invention patents, and 10 of them have been authorized. He has been invited to serve as the chair of subconferences of international academic conferences for many times. He has obtained more than 10 software copyrights. He has published a textbook named Artificial Intelligence and Robotics. He has won one third prize of the Jiangsu Science and Technology Award. He has guided students to participate in competitions and won one first prize and one second prize in provincial competitions, and one third prize in national competitions, etc.

Xinnan Fan, PhD, is a Professor and Doctoral Supervisor at Hohai University. His main research areas include information acquisition and processing, intelligent sensing technology, Internet of Things (IoT) technology and applications, and water conservancy informatization. He has led or participated in the completion of more than 30 projects, including those funded by the National Natural Science Foundation of China, the National "863" Program, and sub-projects of the National Key R&D Program. He has published over 100 academic papers (more than 30 of which are SCI-indexed), obtained 18 authorized invention patents, registered 10 software copyrights, authored one monograph, and received two provincial and ministerial-level science and technology awards, as well as four Jiangsu Provincial Outstanding Teaching Achievement Awards. He was also awarded the Baosteel Education Award for Outstanding Teachers. Currently, he serves as the Director of the Jiangsu Provincial Key Laboratory of Transmission and Distribution Equipment Technology, Deputy Secretary-General of the Jiangsu Transmission and Distribution Industry Technology Innovation Alliance, Outstanding Mid-Career Leader of the Jiangsu "Qinglan Project," and Head of the Water IoT and Sensing Center Team at the Jiangsu "World Water Valley" and Water Ecological Civilization Collaborative Innovation Center. He teaches undergraduate courses such as "Principles and Applications of Single-Chip Microcomputers."

Yuanxue Xin, PhD, is an Associate Professor and Master's Supervisor at Hohai University, as well as a member of the China Computer Federation (CCF) and has been recognized as a "Dayu Scholar" by Hohai University and a "Longcheng Talented Professional" in Changzhou. Her primary research focuses on spectral efficiency, energy efficiency and novel duplexing technologies in massive MIMO systems. She has presided over one National Natural Science Foundation of China (Youth Program) project and one Open Fund from the State Key Laboratory of Mobile Communications. As the first author, she has published 14 papers, including 7 SCI-indexed articles, four EI-indexed articles, and one top-tier conference paper in the field of communications. She has also obtained one authorized invention patent as the first inventor, with two additional patent applications under review. Additionally, she has participated in three international academic conferences and delivered one keynote presentation.

Gang Wan is a Senior Engineer at China Yangtze Power Co., Ltd. His main research areas include intelligent operation and maintenance of large hydropower stations, as well as intelligent management and digitalization of hydropower maintenance processes. He has participated in two national key research and development projects, serving as the lead of one specialized topic. He has received multiple awards, including first and third prizes for outstanding employee technological innovation achievements from organizations such as the China Electricity Council and the National Energy, Chemical, and Geological Systems. He has been granted over 50 patents, including 16 invention patents, and has published 12 papers, three of which are SCI-indexed. In the past three years, he has delivered two presentations at academic conferences.

Qingying Wang is a Senior Engineer at the Beijing Institute of Space Mechanics & Electricity. She has long been engaged in theoretical research, system design, and engineering development in the field of space remote sensing, achieving a series of innovative results in the structural design, simulation, testing, and experimentation of space cameras. She has participated in over ten National Major Science and Technology Projects, serving as the lead of two specialized topics. Her awards include the Third Prize of the National Defense Technology Invention Award, the Third Prize of the Corporate-level Science and Technology Invention Award, as well as multiple institute-level honors in research management, quality work, and "Five Small" outstanding innovation achievements. She has been granted more than ten patents and has published several papers in core journals and academic conferences. In the past three years, she has delivered one presentation at an academic conference.