Article
Computer Science, Information Systems
Yuanfan Yu, Sixian Chan, Tinglong Tang, Xiaolong Zhou, Yuan Yao, Hongkai Zhang
Summary: In the manufacturing process of industrial robots, defect detection of raw materials plays a crucial role in ensuring accuracy. However, existing methods suffer from low precision and generalization ability. This study proposes a detection method using an attention mechanism and dilated convolution module to overcome these disadvantages. By applying a two-stage detection framework with Resnet50 as the pre-training network, the proposed method achieves better results by focusing on effective regions and increasing the model's receptive field through dilated convolution.
Article
Chemistry, Analytical
Khaled R. R. Ahmed
Summary: This paper proposes and develops the DSTEELNet convolution neural network (CNN) architecture to improve the accuracy and efficiency of defect detection in surface steel strips. DSTEELNet consists of three parallel stacks of convolution blocks with atrous spatial pyramid pooling, which expand the receptive fields, enhance the feature resolutions, and cover square regions of input images without any gaps or missing edges. Experimental results demonstrate that DSTEELNet achieves a mean average precision (mAP) of 97% in detecting defects on the augmented dataset GNEU and Severstal datasets, and is capable of detecting defects in a single image within 23ms.
Article
Computer Science, Information Systems
Xiliang Tong, Zhenyu Zhu, Dongxu Bai, Bo Tao, Juntong Yun, Xin Liu, Guojun Zhao
Summary: The new generation of industrial productions is characterized by being digitized, adaptable, flexible, and intelligent. Product inspection now utilizes multiple sources of information and intelligent algorithms for anomaly detection. Cyber-physical systems (CPSs) play a crucial role in smart production and inspection. This article presents a novel approach for solving the challenges in industrial surface inspection by introducing a lighting scheme, multichannel image capture, image splitting, and intelligent classification. The proposed scheme, which combines multichannel photometric images and a classification model, outperforms traditional methods in comparative experiments.
IEEE SYSTEMS JOURNAL
(2023)
Article
Engineering, Electrical & Electronic
Hanshen Chen, Huiping Lin
Summary: Automated pixel-level crack detection is crucial in defect inspection, and deep convolutional neural networks have been successfully applied. However, traditional encoder-decoder networks may have issues with downsampling, upsampling, and a large number of parameters. To address this, a hybrid atrous convolutional network (HACNet) was proposed, maintaining spatial resolution and achieving accurate segmentation with relatively few parameters. The evaluation results demonstrated the promising performance of HACNet compared to recent approaches.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Automation & Control Systems
Wenbo Jiang, Min Liu, Yunuo Peng, Lehui Wu, Yaonan Wang
Summary: The study proposed HDCB-Net for pixel-level detection of blurred cracks, achieving efficient fast crack detection through a two-stage strategy with a processing time of only 0.64 seconds per image. Adding to that, a public dataset comprising 150,632 high-resolution images was established for crack detection research purposes.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Mechanics
Mohammad Bahonar, Mir Saeed Safizadeh
Summary: This research utilized air-coupled ultrasonic testing to detect composite defects, improving delamination detection accuracy through proper test setup and the use of a support vector machine classifier.
COMPOSITE STRUCTURES
(2022)
Article
Computer Science, Information Systems
Zhanzhi Su, Mingle Zhou, Honglin Wan, Min Li, Zekai Zhang, Delong Han, Rui Shao, Gang Li
Summary: This paper proposes an Interactive Convolutional Transformer-based Encoder-Decoder Defect Detection Network (ICT-EDNet) to solve the difficulties in defect detection, by enhancing the semantic information of key features and defects through specific modules and mechanisms.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2023)
Article
Automation & Control Systems
Xiaoqing Zheng, Song Zheng, Yaguang Kong, Jie Chen
Summary: This paper presents the latest advances in surface defect inspection using deep learning, focusing on industrial products in semiconductor, steel, and fabric manufacturing processes. The research provides literature reviews, traditional surface defect inspection algorithms, and applications of deep learning-based inspection algorithms.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Ying Liang, Ke Xu, Peng Zhou, Dongdong Zhou
Summary: In this article, a simple yet powerful image transformation network is proposed to remove textures and highlight defects at full resolution. The network utilizes a polynomial loss function combining perceptual loss, structural similarity loss, and image gradient loss to effectively suppress texture and emphasize defects. The method demonstrates superior performance in experiments and has been successfully applied to the surface defect online detection system of an aluminum ingot milling production line.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Chemistry, Multidisciplinary
Hongjun Wang, Xiujin Xu, Yuping Liu, Deda Lu, Bingqiang Liang, Yunchao Tang
Summary: Due to surface defects, inadequate contrast, and resemblance between noise and defects, edge detection is challenging in dimensional error detection. This research proposes a combined approach using the YOLOv6 deep learning network and metal lock body parts for accurate and rapid defect detection. An enhanced Canny-Devernay algorithm is used for size measurement.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Yijing Guo, Yixin Zeng, Fengqiang Gao, Yi Qiu, Xuqiang Zhou, Linwei Zhong, Choujun Zhan
Summary: This paper proposes an improved algorithm based on YOLOV4-CSP for bamboo surface defect detection. The introduction of asymmetric convolution and attention mechanism enhances the detection performance of specific industrial defects. Experimental results demonstrate the outstanding performance of the algorithm in general and hard-to-detect categories.
Review
Computer Science, Information Systems
Wuyi Ming, Chen Cao, Guojun Zhang, Hongmei Zhang, Fei Zhang, Zhiwen Jiang, Jie Yuan
Summary: With the rapid development of 3C products, the demand for quality display panels and related detection technologies is increasing. Convolutional Neural Networks play a key role in defect detection, and this article summarizes the methods and applications of CNN with different depths in 3C product defect detection.
Article
Automation & Control Systems
Weiping Ding, Janmenjoy Nayak, Bighnaraj Naik, Danilo Pelusi, Manohar Mishra
Summary: The article introduces a cluster analysis approach for intrusion detection system in Big Data platform, which utilizes feature selection to avoid processing a large number of features, thereby enhancing the model's efficiency.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Computer Science, Artificial Intelligence
Qianru Zhang, Meng Zhang, Chinthaka Gamanayake, Chau Yuen, Zehao Geng, Hirunima Jayasekara, Chia-wei Woo, Jenny Low, Xiang Liu, Yong Liang Guan
Summary: With the improvement of electronic circuit production methods, the risk of defects in the production line is increasing. X-ray imaging is able to inspect defects, but current algorithms are not accurate enough. This study introduces deep learning into quality control inspection, proposes four joint defect detection models and addresses the issues of noisy regions of interest and changing image dimensions.
COMPLEX & INTELLIGENT SYSTEMS
(2022)
Article
Computer Science, Theory & Methods
Eman Badr
Summary: Medical imaging diagnosis relies heavily on medical experts, but with the advancement of deep learning techniques, new methods have been developed to assist in the diagnosis and risk assessment processes. Radiomics concepts and research methods show potential for integrated diagnostics, while convolutional neural networks play a key role in automated imaging biomarker extraction and big data analytics systems for future generations.
ACM COMPUTING SURVEYS
(2021)
Article
Computer Science, Information Systems
Xiaojie Wang, Laisen Nie, Zhaolong Ning, Lei Guo, Guoyin Wang, Xinbo Gao, Neeraj Kumar
Summary: This article studies the problem of end-to-end network traffic prediction in the backbone networks of Internet of Vehicles (IoV), and proposes a deep learning-based method which considers the spatio-temporal feature and long-range dependence of network traffic. Furthermore, a threshold-based update mechanism is introduced to improve the real-time performance of the method.
ACM TRANSACTIONS ON INTERNET TECHNOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Zhaolong Ning, Handi Chen, Xiaojie Wang, Shupeng Wang, Lei Guo
Summary: The maturity of 5G communication technology promotes industrial revolution and high-quality development of the economy. However, remote grid inspection and maintenance face challenges due to scarce resources and complex environments. To solve this, a blockchain-enabled secure transmission scheme and improved market matching algorithm are proposed, with deep reinforcement learning and A* algorithm used for secure transmission.
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING
(2022)
Article
Engineering, Electrical & Electronic
Zeng-qi Li, Xin Liu, Zhao-long Ning
Summary: In this paper, a DSA scheme based on deep reinforcement learning combined with multiple access methods is proposed to maximize system throughput. The scheme intelligently learns the best access strategy in a dynamic environment and avoids interference with other users.
PHYSICAL COMMUNICATION
(2022)
Article
Engineering, Multidisciplinary
Xiaoqiang Zhu, Wenyu Qu, Xiaobo Zhou, Laiping Zhao, Zhaolong Ning, Tie Qiu
Summary: In this paper, a novel intelligence localization scheme called ILCL is proposed, which utilizes incremental learning and deep neural networks to improve the positioning accuracy and reduce the time-consuming retraining. Experimental results in two real indoor environments confirm the superiority of ILCL.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2022)
Article
Automation & Control Systems
Handi Chen, Shupeng Wang, Guojun Li, Laisen Nie, Xiaojie Wang, Zhaolong Ning
Summary: This article establishes a dynamic network virtualization technique enabled service function chain orchestration framework in IIoT and proposes a dynamic orchestration scheme, which is validated to be superior through experimental results.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Tie Qiu, Xize Liu, Xiaobo Zhou, Wenyu Qu, Zhaolong Ning, C. L. Philip Chen
Summary: In this paper, an adaptive social spammer detection (ASSD) model is proposed to effectively identify spammers in mobile social networks. The model achieves high accuracy and efficiency, and updates adaptively through incremental learning.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2022)
Article
Computer Science, Information Systems
Xiaojie Wang, Zhaolong Ning, Song Guo, Miaowen Wen, H. Vincent Poor
Summary: This paper focuses on the minimization of Age-of-Critical-Information (AoCI) in mobile edge networks. A system model is established to quantify the freshness of information by changes in its critical levels. An information-aware heuristic algorithm and an imitation learning-based scheduling approach are proposed to address the problem, and the superiority of the designed algorithm is demonstrated from both theoretical and experimental perspectives.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)
Article
Computer Science, Information Systems
Zhaolong Ning, Shouming Sun, Xiaojie Wang, Lei Guo, Song Guo, Xiping Hu, Bin Hu, Ricky Y. K. Kwok
Summary: Intelligent Transportation System (ITS) is essential for addressing traffic issues and providing services for personal travel. However, existing research has not comprehensively considered the safety, utility, and latency of user data. Therefore, we propose a blockchain-enabled crowdsensing framework for distributed traffic management. By using two algorithms, we achieve secure and efficient transmissions, and experimental results demonstrate the effectiveness of these algorithms in improving the performance of the transportation system.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)
Article
Computer Science, Information Systems
Yixuan Wu, Laisen Nie, Shupeng Wang, Zhaolong Ning, Shengtao Li
Summary: With the rapid growth of IoT, cloud-centric computing struggles to meet the low latency and ease of use requirements. Edge computing, as an open and distributed system, integrates computing, networking, storage, and applications, providing intelligent services at the IoT edge. However, the edge network faces various cyber attacks due to its limited resources, making large-scale data collection and detection for IoT security challenging. This paper proposes an intelligent intrusion detection algorithm based on big data mining and a combination of fuzzy rough set, GAN, and CNN, achieving higher accuracy than existing methods.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Editorial Material
Computer Science, Information Systems
Uttam Ghosh, Deepak Tosh, Nawab Muhammad Faseeh Qureshi, Ali Kashif Bashir, Al-Sakib Khan Pathan, Zhaolong Ning
Article
Computer Science, Information Systems
Yuhuai Peng, Zhaolong Ning, Aiping Tan, Shupeng Wang, Mohammad S. Obaidat
Summary: This paper proposes a delay-sensitive multibase-station multichannel access scheme to address the real-time transmission problem in IIoT systems. Experimental results show that compared to CSMA/CA, this scheme can significantly reduce packet loss and average latency.
IEEE SYSTEMS JOURNAL
(2023)
Article
Computer Science, Information Systems
Xiaojie Wang, Zhaolong Ning, Song Guo, Miaowen Wen, Lei Guo, H. Vincent Poor
Summary: This study proposes a multi-agent imitation learning enabled UAV deployment approach to enable different UAV owners to provide services with differentiated service capabilities in a shared area. The goal is to maximize both profits of UAV owners and utilities of on-ground users.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Computer Science, Information Systems
Zhaolong Ning, Yuxuan Yang, Xiaojie Wang, Lei Guo, Xinbo Gao, Song Guo, Guoyin Wang
Summary: In this paper, a MEC network enabled by UAVs is investigated, considering multi-user computation offloading and edge server deployment to minimize system-wide computation cost under a dynamic environment. The problem is decomposed into two stochastic games and it is proven that each game has at least one Nash Equilibrium. Two learning algorithms are proposed to reach the Nash Equilibriums. These algorithms are further incorporated into an asynchronous updating algorithm to solve the system-wide computation cost minimization problem. Performance evaluations based on real-world data are conducted, showing the proposed algorithms can achieve efficient computation offloading and server deployment under dynamic environments.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Proceedings Paper
Computer Science, Information Systems
Jian Song, Hua Liu, Laisen Nie, Zhaolong Ning, Mohammad S. Obaidat, Balqies Sadoun
Summary: This paper proposes a novel algorithm that combines Deep Q-Learning and Generative Adversarial Networks for network traffic prediction, and the performance of this algorithm is proven to be better than other methods through experiments.
2022 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM 2022)
(2022)
Article
Engineering, Civil
Guangjun Wu, Jun Li, Zhaolong Ning, Yong Wang, Binbin Li
Summary: This paper investigates the task scheduling and running of transportation big data (TBD) based on credit priority. A three-layered architecture is designed, and a federated learning mechanism is proposed to protect the privacy of vehicles and improve the efficiency of task scheduling. A task offloading algorithm is also proposed to optimize the task offloading problem.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
