Content of Research Hotspots and Reviews in our journal

Published in last 1 year |  In last 2 years |  In last 3 years |  All Please wait a minute... For Selected: Download Citations EndNote Ris BibTeX Toggle Thumbnails Select Review of Application of Visual Foundation Model SAM in Medical Image Segmentation SUN Xing, CAI Xiaohong, LI Ming, ZHANG Shuai, MA Jingang Computer Engineering and Applications    204, 60 (17): 1-16.   DOI: 10.3778/j.issn.1002-8331.2401-0136 Abstract （87）      PDF（pc） （7912KB）（94）       Knowledge map    Save With the continuous development of foundation models technology, visual foundation model represented by the segment anything model (SAM) has made significant breakthroughs in the field of image segmentation. SAM, driven by prompts, accomplishes a series of downstream segmentation tasks, aiming to address all image segmentation issues comprehensively. Therefore, the application of SAM in medical image segmentation is of great significance, as its generalization performance can adapt to various medical images, providing healthcare professionals with a more comprehensive understanding of anatomical structures and pathological information. This paper introduces commonly used datasets for image segmentation, provides detailed explanations of SAM’s network architecture and generalization capabilities. It focuses on a thorough analysis of SAM’s application in five major categories of medical images: whole-slide imaging, magnetic resonance imaging, computed tomography, ultrasound, and multimodal images. The review summarizes the strengths and weaknesses of SAM, along with corresponding improvement methods. Combining current challenges in the field of medical image segmentation, the paper discusses and anticipates future directions for SAM’s development. Reference | Related Articles | Metrics Select Comprehensive Review of Large Language Model Fine-Tuning ZHANG Qintong, WANG Yuchao, WANG Hexi, WANG Junxin, CHEN Hai Computer Engineering and Applications    204, 60 (17): 17-33.   DOI: 10.3778/j.issn.1002-8331.2312-0035 Abstract （52）      PDF（pc） （6335KB）（67）       Knowledge map    Save The rise of large-scale language models signifies a new milestone in the field of deep learning, with fine-tuning techniques playing a crucial role in optimizing model performance. This paper provides a comprehensive overview of fine-tuning techniques for large-scale language models. It reviews the development stages of language models, including statistical language models, neural network language models, pre-trained language models, and large language models. The basic concepts of fine-tuning are explored, covering classic fine-tuning, efficient parameter fine-tuning, prompt tuning, and reinforcement learning fine-tuning. The paper delves into the principles and development of each fine-tuning technique, offering a comparative analysis across these four major categories. In conclusion, the paper summarizes the current state of research on fine-tuning techniques and underscores the potential research value in this domain, providing insights into future directions of development. Reference | Related Articles | Metrics Select Comprehensive Review of ROV Underwater Obstacle Detection and Avoidance Technology LI Minggui, ZHOU Huanyin, GONG Liwen Computer Engineering and Applications    204, 60 (17): 34-47.   DOI: 10.3778/j.issn.1002-8331.2312-0206 Abstract （33）      PDF（pc） （5168KB）（27）       Knowledge map    Save This paper provides a comprehensive review of the technological advancements in underwater obstacle detection and avoidance techniques for remotely operated vehicles (ROV). The research focuses on sonar systems, optical systems, and their integration with machine learning and artificial intelligence algorithms, analyzing how these technologies enhance the autonomy, efficiency, and safety of underwater operations. Despite significant achievements in environmental adaptability and obstacle detection accuracy achieved by sonar and optical systems, challenges remain in real-time identification of dynamic obstacles and adaptation to complex environments. Furthermore, the potential and challenges of machine learning and artificial intelligence technologies in enhancing ROV’s autonomous obstacle avoidance capability are discussed, highlighting the importance of these technologies in future ROV operations. This research provides new theoretical perspectives and practical applications for deep-sea exploration and marine science. Reference | Related Articles | Metrics Select Review of Frequent Temporal Pattern Mining Methods TANG Zengjin, XU Zhenshun, SU Mengyao, LIU Na, WANG Zhenbiao, ZHANG Wenhao Computer Engineering and Applications    204, 60 (17): 48-61.   DOI: 10.3778/j.issn.1002-8331.2403-0114 Abstract （23）      PDF（pc） （5060KB）（28）       Knowledge map    Save Frequent temporal pattern mining refers to the process of discovering frequently occurring patterns or patterns from time series data. Its purpose is to help understand important features in time series data, such as periodicity, trends, and anomalies, which can help predict future development trends and identify abnormal situations. Based on literature research on frequent temporal pattern mining methods in recent years, they are divided into three categories according to key technologies and representative algorithms, namely structural constraint based frequent temporal pattern mining methods, parameter constraint based frequent temporal pattern mining methods, and window based frequent temporal pattern mining methods. Firstly, the background of frequent temporal pattern mining methods and the characteristics of each method are described. Secondly, the development and classification of three mining methods are introduced, and a detailed comparative analysis is conducted on the advantages, disadvantages, and performance of each improved method. Finally, the frequent temporal pattern mining methods are summarized and summarized, and the future research directions of frequent temporal pattern mining methods are discussed. Reference | Related Articles | Metrics Select Overview of GNSS Spoofing Detection Using Machine Learning ZHOU Yalan, SONG Xiao’ou Computer Engineering and Applications    204, 60 (17): 62-73.   DOI: 10.3778/j.issn.1002-8331.2401-0330 Abstract （26）      PDF（pc） （4308KB）（27）       Knowledge map    Save In recent years, with the increasing importance of global satellite navigation, deception detection has become a hot research issue. As a low-cost method, machine learning has the ability to automatically learn rules from complex data, and has achieved good results in the Internet of things spoofing detection. Therefore, more and more studies have used it for GNSS spoofing detection. Firstly, starting from the basic process of GNSS spoofing detection based on machine learning, the data acquisition and preprocessing of GNSS detection using machine learning are described. Then, according to the role of machine learning in spoofing detection, GNSS spoofing detection based on machine learning is divided into two categories: GNSS spoofing detection based on signal classification and GNSS spoofing detection based on information verification. Finally, according to the existing research problems, the prospect of future development direction is put forward. Reference | Related Articles | Metrics Select Review of Research on Artificial Intelligence in Traditional Chinese Medicine Diagnosis and Treatment SU Youli, HU Xuanyu, MA Shijie, ZHANG Yuning, Abudukelimu Abulizi, Halidanmu Abudukelimu Computer Engineering and Applications    2024, 60 (16): 1-18.   DOI: 10.3778/j.issn.1002-8331.2312-0400 Abstract （195）      PDF（pc） （6171KB）（215）       Knowledge map    Save The field of traditional Chinese medicine (TCM) diagnosis and treatment is gradually moving towards standardization, objectification, modernization, and intelligence. In this process, the integration of artificial intelligence (AI) has greatly propelled the advancement of TCM diagnosis and treatment, scientific research, and TCM inheritance. The review starts from the current research status of AI in TCM, combs through the application and development of AI in TCM in three stages from expert system and rule engines, traditional machine learning algorithm to deep learning, and then summarizes the knowledge management tools and large language models of TCM in recent years. Finally, this paper analyzes the multiple challenges of data fairness, multimodal data understanding, model robustness, personalized medicine, and interpretability that exist at this stage of AI in TCM. To address these challenges, it is necessary to continuously explore and propose possible solutions to promote the in-depth development of intelligent TCM diagnosis and treatment, thus better meeting the health needs of people. Reference | Related Articles | Metrics Select Survey of Pattern Mining Methods Based on Biological Heuristic Algorithms HAN Meng, HE Feifei, ZHANG Ruihua, LI Chunpeng, MENG Fanxing Computer Engineering and Applications    2024, 60 (16): 19-33.   DOI: 10.3778/j.issn.1002-8331.2401-0427 Abstract （87）      PDF（pc） （5424KB）（86）       Knowledge map    Save Frequent itemset mining, association rule mining and high utility itemset mining are three related and developing fields in pattern mining. In recent years, because traditional algorithms cannot cope with the explosive growth of data environment, heuristic algorithms have become a research hotspot in pattern mining methods. In order to reveal the research and development status in the field of pattern mining, firstly, the research results of frequent patterns and high utility pattern methods are comprehensively analyzed and summarized from the perspective of many biological heuristic algorithms, such as particle swarm optimization, genetic algorithm, ant colony optimization, and artificial bee colony and so on. Secondly, different biological heuristic pattern mining methods are summarized from strategy, comparison algorithm, datasets, advantages and disadvantages, and the experimental results and performance indicators of the same datasets are compared and analyzed in detail. Finally, in view of the shortcomings of the current biological heuristic pattern mining methods, the next research direction is put forward, including dynamic data flow, multi-objective evolution, fuzzy computing and complex data types. Reference | Related Articles | Metrics Select Review of Data-Driven Approaches to Chinese Named Entity Recognition XIAO Lei, CHEN Zhenjia Computer Engineering and Applications    2024, 60 (16): 34-48.   DOI: 10.3778/j.issn.1002-8331.2312-0260 Abstract （70）      PDF（pc） （5433KB）（65）       Knowledge map    Save Chinese named entity recognition (CNER) is a key step in Chinese information extraction task, which is the basis of downstream tasks such as question answering system, machine translation and knowledge mapping, and its methods are mainly categorized into two main types: knowledge-driven and data-driven. However, the traditional knowledge-driven methods based on rules, dictionaries and machine learning have the problems of ignoring contextual semantic information, high computational cost and low recall rate, which limit the development of CNER technology. Firstly, the definition and development history of CNER are introduced. Secondly, the typical datasets, training tools, sequence annotation methods and model evaluation indexes for CNER tasks are organized in detail. Thirdly, the data-driven methods are summarized and divided into methods based on deep learning, pre-trained language models and joint extraction of Chinese entity relations, and the practical application scenarios of data-driven methods in different fields are analyzed. Finally, the future research direction of CNER task is outlooked to provide some reference for the proposal of new methods. Reference | Related Articles | Metrics Select Review of Research on Lightweight Image Super-Resolution ZHU Xinfeng, SONG Jian Computer Engineering and Applications    2024, 60 (16): 49-60.   DOI: 10.3778/j.issn.1002-8331.2403-0230 Abstract （66）      PDF（pc） （4762KB）（66）       Knowledge map    Save In recent years, image super-resolution (SR) based on deep learning has received widespread attention. The purpose of image SR is to improve the resolution of images to facilitate further processing of images, such as target detection, image classification and face recognition, etc. The research on image SR has achieved rapid development in recent years, but there are still few related reviews on lightweight SR models. By analyzing the current research status of lightweight SR methods which are based on deep learning and loss function, a new classification of current lightweight SR models is made, which are traditional convolution methods and attention mechanism methods. The development history and latest progress of lightweight SR methods for images are systematically given, the advantages and disadvantages of each method are pointed out. Finally, by analyzing the existing problems of current lightweight SR technology, the future research directions of lightweight image SR method are given. Reference | Related Articles | Metrics Select Review of Deep Learning Algorithms for One-Stage Safety Helmet Detection GUAN Hanyu, LING Yun, WANG Shulei Computer Engineering and Applications    2024, 60 (16): 61-75.   DOI: 10.3778/j.issn.1002-8331.2312-0034 Abstract （59）      PDF（pc） （5273KB）（62）       Knowledge map    Save Real-time detection of safety helmet-wearing is an essential part of smart construction sites and smart traffic, safety helmet detection based on deep learning has gradually replaced the traditional detection methods, and has made significant progress in accuracy, performance, and efficiency. It has been widely used in real scenarios. To facilitate the future research of safety helmet-wearing algorithms, the research status of object detection algorithms for safety helmets in various application scenarios is analyzed comprehensively. Firstly, the history of the object detection algorithm is summarized. Secondly, the advantages and disadvantages of different algorithms and optimizations are analyzed, and the lightweight safety helmet detection algorithms are discussed. Finally, according to the shortages of the current object detection algorithm applied in the actual scene, the future direction of a deep learning algorithm for safety helmet detection is prospected. Reference | Related Articles | Metrics Select Research on Unmanned Aerial Vehicle Swarm Resilience Assessment and Reconfiguration Technology WEI Chenyue, HE Ming, HAN Wei, XU Xin, GAO Hong Computer Engineering and Applications    2024, 60 (15): 1-10.   DOI: 10.3778/j.issn.1002-8331.2401-0452 Abstract （154）      PDF（pc） （4418KB）（203）       Knowledge map    Save Unmanned aircraft vehicle (UAV) swarm is often affected by perturbing factors such as terrain, wind, snow, rain and fog, and anti-aircraft strikes in practical applications, which leads to the decline of swarm performance and mission accomplishment capability. In order to effectively assess and improve the swarm anti-disturbance capability, an in-depth study is carried out in terms of UAV swarm resilience assessment indexes and resilience reconfiguration methods. Firstly, the current research status of UAV swarm resilience assessment indicators is sorted out and analyzed. Secondly, the research on UAV swarm resilience reconstruction methods is summarized in terms of predictive reconstruction and anti-disturbance reconstruction. To address the problems of incomplete assessment indexes and the inability of swarm adaptive reconfiguration under multi-task and multi-disturbance situations, multi-dimensional resilience assessment indexes and UAV swarm phase change reconfiguration methods are proposed respectively, which further take into account the impact of coverage, energy consumption and other factors on swarm performance, realize the adaptive phase change of different types of tasks and disturbance types, and significantly improve the swarm’s ability to cope with disturbances. Finally, it concludes and looks forward to the future development trend of UAV swarm elastic reconfiguration. Reference | Related Articles | Metrics Select Comprehensive Review of Multimodal Data Fusion Methods for Construction Robot Localization LI Jiayi, MA Zhiliang, CHENG Lijie, JI Xinlin Computer Engineering and Applications    2024, 60 (15): 11-23.   DOI: 10.3778/j.issn.1002-8331.2402-0008 Abstract （83）      PDF（pc） （4423KB）（101）       Knowledge map    Save The types of localization data sources for construction robots are diverse, and the fusion of multimodal data not only helps to improve the localization performance of construction robots in building projects but also facilitates their collaborative operations. Data fusion methods aim to enhance construction robot localization and data sharing by leveraging the advantages of different data sources, improving data collection and processing methods, this supports the improvement of construction robot localization accuracy, real-time performance, and robustness, ultimately enhancing overall construction efficiency and project management. While there are existing research outcomes on specific scenarios exploring data fusion methods for construction robot positioning, there is currently no comprehensive review article on this topic. Through systematic retrieval, this paper first categorizes the analysis into two types based on whether it integrates with prior data：fusion of prior data and real-time sensor data fusion, and fusion of data from multiple sensors. Subsequently, a comparative analysis of data fusion methods is conducted. Finally, the paper summarizes and anticipates the future research directions of multimodal data fusion methods for construction robots. The analysis of current research results reveals significant variations in the choice and effectiveness of positioning data sources for construction robots. This review can serve as a reference for further research in related fields. Reference | Related Articles | Metrics Select Review of Research on Natural Language Interfaces for Data Visualization GAO Shuai, XI Xuefeng, ZHENG Qian, CUI Zhiming, SHENG Shengli Computer Engineering and Applications    2024, 60 (15): 24-41.   DOI: 10.3778/j.issn.1002-8331.2310-0167 Abstract （83）      PDF（pc） （7584KB）（121）       Knowledge map    Save The long-standing goal in the field of data visualization has been to find a solution that directly generates visualizations from natural language. Research on natural language interfaces (NLI) provides a new approach to this field. This interface accepts queries in the form of natural language and tabular datasets as input and outputs corresponding visualization renderings. Simultaneously, as an auxiliary input method, traditional users need to convert analytical intents into a series of logical operations and interact with them, such as programming instructions or graphical interface operations. Combining the use of natural language interfaces for data visualization (DV-NLI) enables users to focus on visualization tasks without worrying about how to operate visualization tools. In recent years, with the rise of large language models (LLM) such as GPT-3 and GPT-4, research on integrating LLM with visualization has become a hot topic. This paper provides a comprehensive review of existing DV-NLIs and supplements them with the latest research. Based on their implementation methods, DV-NLIs are categorized into symbolic NLP methods, deep learning model methods, and large language model methods. It also analyzes and discusses relevant techniques under each category. Finally, the paper summarizes and looks forward to future work in DV-NLI. Reference | Related Articles | Metrics Select Review on Zero or Few Sample Rotating Machinery Fault Diagnosis LIU Junfu, CEN Jian, HUANG Hankun, LIU Xi, ZHAO Bichuang, SI Weiwei Computer Engineering and Applications    2024, 60 (15): 42-54.   DOI: 10.3778/j.issn.1002-8331.2401-0112 Abstract （34）      PDF（pc） （5599KB）（39）       Knowledge map    Save With the advent of the data era, data-driven fault diagnosis methods have demonstrated excellent performance. Since the application of deep learning in fault diagnosis, supervised learning has made significant advancements. However, when samples are scarce or missing, supervised learning lacks the necessary training conditions. This paper proposes the zero-shot and small-sample problem, and analyzes its current status in the field of rotating machinery fault diagnosis. It reviews the development process, mainstream models, and current research hotspots of zero-shot rotating machinery fault diagnosis. Existing research achievements are summarized from two aspects: zero-shot problems and small-sample problems, and their applications in zero-shot and small-sample problems are analyzed. Finally, the paper discusses the future trends in zero-shot methods for rotating machinery fault diagnosis. Reference | Related Articles | Metrics Select Review of Image Denoising Methods for Remote Sensing WANG Haoyu, YANG Haitao, WANG Jinyu, ZHOU Xixuan, ZHANG Honggang, XU Yifan Computer Engineering and Applications    2024, 60 (15): 55-65.   DOI: 10.3778/j.issn.1002-8331.2402-0190 Abstract （55）      PDF（pc） （13356KB）（64）       Knowledge map    Save The complexity of the imaging environment results in remote sensing images containing many types of noise, and the removal of these noises can effectively improve the efficiency and accuracy of the subsequent work. In recent years, image denoising methods for remote sensing have gradually become a hotspot in the field of image processing. On the basis of absorbing the research of many scholars at home and abroad, the denoising methods for visible remote sensing images, infrared remote sensing images and SAR images are systematically summarized. Firstly, the main sources and manifestations of noise in remote sensing images are introduced. Secondly, public datasets and data platforms that can be used for the study of remote sensing image denoising methods are listed. The advantages and limitations of traditional remote sensing image denoising methods are described according to the processing domain. Then, the cutting-edge image denoising method for remote sensing based on deep learning is highlighted, and its main innovations and shortcomings are summarized. Finally, the challenges and future development directions of remote sensing image denoising task are analyzed and prospected. Reference | Related Articles | Metrics Select Research Review on Deep Reinforcement Learning for Solving End-to-End Navigation Problems of Mobile Robots HE Li, YAO Jiacheng, LIAO Yuxin, ZHANG Wenzhi, LU Zhaoqing, YUAN Liang, XIAO Wendong Computer Engineering and Applications    2024, 60 (14): 1-13.   DOI: 10.3778/j.issn.1002-8331.2312-0256 Abstract （155）      PDF（pc） （4646KB）（315）       Knowledge map    Save Autonomous navigation is the prerequisite and foundation for mobile robots to accomplish complex tasks. Traditional autonomous navigation systems rely on the accuracy of maps and cannot adapt to highly complex industrial and service scenarios. End-to-end navigation methods for mobile robots that do not rely on a priori map information and are able to make autonomous decisions through deep reinforcement learning, and environment interaction learning have become a new research hotspot. Most existing classifications cannot comprehensively summarize the challenges and opportunities of end-to-end navigation problems. Based on the characteristics of end-to-end navigation systems, the challenges of the navigation problem are attributed to the key issues of poor perception ability of navigation agents, ineffective learning and poor generalization ability of navigation strategies. The research status and development trends of end-to-end navigation systems are described. Representative research results in recent years addressing these key issues are detailed respectively, and their advantages and shortcomings are summarized. Finally, the future development trends of end-to-end navigation for mobile robots are prospectively envisioned in aspects such as visual language navigation, multi-agents collaborative navigation, end-to-end navigation for fusion super-resolution reconstructed images and interpretable end-to-end navigation, providing certain insights for the research and application of end-to-end navigation for mobile robots. Reference | Related Articles | Metrics Select Survey on Recent Advances in Context Awareness of Augmented Reality YANG Zhuo, MAI Eryuan, LI Huicong, MO Jianqing Computer Engineering and Applications    2024, 60 (14): 14-25.   DOI: 10.3778/j.issn.1002-8331.2312-0227 Abstract （24）      PDF（pc） （4808KB）（15）       Knowledge map    Save Augmented reality provides users with immersive digital experiences by overlaying digital content on top of physical world and helps users better understand reality. In recent years，research on context awareness in augmented reality has received widespread attention. Augmented reality is closely related to its usage context，not only it is basic contextual information such as location information, light level, head posture, but also higher level contextual information such as geometric information, semantic information and eye movement information, these contextual information have impacts on user experience of augmented reality in various degrees. Based on recent advances in context awareness of augmented reality, this paper outlines the contextual factors of augmented reality, illustrates the context-aware processing flow in augmented reality, and analyzes the context-aware algorithms and latest applications related to geometry, semantics and eye movement contextual information in augmented reality. Finally, combined with the current status of augmented reality contextual awareness, it discusses the future development trends of augmented reality contextual awareness, so as to provide reference for future research. Reference | Related Articles | Metrics Select Research Progress in Surface Electromyography Joint Continuous Motion Estimation MA Yifan, WEI Dejian, FENG Yanyan, YU fengfan, LI Zhenjiang Computer Engineering and Applications    2024, 60 (14): 26-36.   DOI: 10.3778/j.issn.1002-8331.2312-0288 Abstract （12）      PDF（pc） （4911KB）（9）       Knowledge map    Save Surface electromyography (sEMG) is a non-invasive bioelectrical signal used to capture changes in muscle activity during exercise. Because it is closely related to sports, it is widely used in the research and development process of intelligent assisted rehabilitation equipment to provide support and assistance for rehabilitation patients. Rehabilitation training involves complex three-dimensional motion, and sEMG-based joint continuous motion estimation is a method to estimate joint angle or moment by analyzing sEMG during exercise, which can effectively alleviate the problem of insufficient adaptability between rehabilitation machine and human body, providing safer assistance and significantly improving the rehabilitation effect. This paper firstly introduces the current status of joint continuous motion estimation, and then classifies the existing sEMG joint continuous motion estimation models into biomechanics-based musculoskeletal model and machine learning-based regression model according to different research methods, and summarizes and analyzes the relevant models respectively. In addition, the paper also analyzes the current challenges and looks forward to the future research trends. Reference | Related Articles | Metrics Select Overview of Research Progress in Graph Transformers ZHOU Chengchen, YU Qiancheng, ZHANG Lisi, HU Zhiyong, ZHAO Mingzhi Computer Engineering and Applications    2024, 60 (14): 37-49.   DOI: 10.3778/j.issn.1002-8331.2308-0216 Abstract （50）      PDF（pc） （4904KB）（72）       Knowledge map    Save With the widespread application of graph structured data in various practical scenarios, the demand for effective modeling and processing is increasing. Graph Transformers (GTs), as a type of model that uses Transformers to process graph data, can effectively alleviate the problems of over smoothing and over squeezing in traditional graph neural network (GNN), and thus can learn better feature representations. Firstly, based on the research on recent GTs related literature, the existing model architectures are divided into two categories: the first category adds graph position and structure information to Transformers through absolute encoding and relative encoding to enhance Transformers’ understanding and processing ability of graph structure data; the second type combines GNN with Transformers in different ways (serial, alternating, parallel) to fully utilize their advantages. Secondly, the application of GTs in fields such as information security, drug discovery, and knowledge graphs is introduced, and the advantages and disadvantages of models with different uses are compared and summarized. Finally, the challenges faced by future research on GTs are analyzed from aspects such as scalability, complex graphs, and better integration methods. Reference | Related Articles | Metrics Select Comprehensive Review on Application of Attention Mechanism in Retinal Vessel Segmentation PEI Junpeng, WANG Yousong, LI Zenghui, WANG Wei Computer Engineering and Applications    2024, 60 (14): 50-65.   DOI: 10.3778/j.issn.1002-8331.2311-0049 Abstract （43）      PDF（pc） （7103KB）（41）       Knowledge map    Save Automatic segmentation of retinal vessels plays an important role in computer-aided diagnosis of ophthalmology and cardiovascular diseases. Attention mechanism can improve the efficiency and accuracy of image feature extraction in classical neural network models, so attention mechanism is widely used in retinal vessel segmentation models. This paper firstly reviews the commonly used datasets and evaluation metrics for retinal vessel segmentation, subsequently, attention mechanisms are categorized into two types: selective attention mechanisms and self-attention mechanisms, based on their working principles. Meanwhile, according to the data domain of computer vision tasks, attention methods are divided into three categories: channel attention, spatial attention, and mixed attention. Combined with the task of retinal vessel segmentation, the paper highlights the specific applications of representative attention models of these three types and conducts performance comparisons and evaluations of relevant models. Finally, the problems of attention mechanism and the development trend in the future are discussed. Reference | Related Articles | Metrics Select Survey on Automated Recognition and Extraction of TTPs YU Fengrui Computer Engineering and Applications    2024, 60 (13): 1-22.   DOI: 10.3778/j.issn.1002-8331.2309-0489 Abstract （144）      PDF（pc） （7424KB）（194）       Knowledge map    Save In the ever-evolving landscape of cyber threats, tactics, techniques and procedures (TTPs) play a crucial role in understanding malicious activities, providing a fine-grained perspective on the status of cybersecurity, and comprehensively illustrating cyber attack behaviors. Despite significant research efforts in the field of automated identification and extraction of TTPs, a comprehensive systematic review is currently lacking. This paper presents an in-depth analysis of the progress in this area by employing three principal approaches：traditional natural language processing, machine learning, and large language models. The study categorizes the tasks into information extraction, text classification, and text generation, and presents a summary of the general framework for identification and extraction processes. It offers a clear scope of unstructured text and TTPs, while refining the processing and analysis procedures, as well as innovative directions for each approaches. Moreover, building upon existing research, the paper identifies current challenges and proposes future research directions and development opportunities. This comprehensive survey serves as a valuable literature review to support readers in applying advanced technologies and methods for advancing research in this field. Reference | Related Articles | Metrics Select Review of Object Detection Based on Event Cameras ZHANG Yali, TIAN Qichuan, TANG Chaolin Computer Engineering and Applications    2024, 60 (13): 23-35.   DOI: 10.3778/j.issn.1002-8331.2312-0322 Abstract （128）      PDF（pc） （5613KB）（173）       Knowledge map    Save Event cameras are imaging methods that mimic biological retinas, with high dynamics, low latency, high temporal resolution and low power consumption. It breaks through the dilemma that traditional cameras are difficult to capture objects and target recognition under high dynamic range, and the characteristics of event cameras are of experimental significance for studying the object detection problem based on event cameras. This paper first briefly describes the status, development process, advantages and challenges of event cameras, then introduces the working principle of various types of event cameras and some object detection algorithms based on event cameras, and finally explains the challenges and future trends of object detection algorithms based on event cameras, and summarizes the article. Reference | Related Articles | Metrics Select Research and Comprehensive Review on Multi-Modal Knowledge Graph Fusion Techniques CHEN Youren, LI Yong, WEN Ming, SUN Chi Computer Engineering and Applications    2024, 60 (13): 36-50.   DOI: 10.3778/j.issn.1002-8331.2309-0481 Abstract （85）      PDF（pc） （6082KB）（86）       Knowledge map    Save Multi-modal knowledge graphs (MMKG) integrate various modal information such as vision and text, presenting knowledge structures graphically. With the advancement of artificial intelligence, MMKG have played a significant role in recommendation systems, intelligent Q&A, and knowledge search among other fields. Compared to traditional knowledge graphs, MMKG can understand and present knowledge in multiple dimensions, possessing superior representation and application capabilities. To delve deep into the study of MMKG, this review first conducts a detailed analysis and elucidation of the value and categories of MMKG. Based on different construction methods, it compares and summarizes multi-modal knowledge extraction, representation learning, entity alignment, and other aspects, categorizes multi-modal knowledge integration methods. It analyzes the progress in the applications of MMKG, discusses the limitations of MMKG, and proposes future research directions in the field of MMKG. Reference | Related Articles | Metrics Select Survey of Application of Graph Neural Network in Anomaly Detection CHEN Jiale, CHEN Xu, JING Yongjun, WANG Shuyang Computer Engineering and Applications    2024, 60 (13): 51-65.   DOI: 10.3778/j.issn.1002-8331.2310-0234 Abstract （68）      PDF（pc） （5395KB）（65）       Knowledge map    Save Graph data is commonly used to represent complex relationships between different individuals, such as social networks, financial networks, and microservice networks. Graph neural network (GNN) is a deep learning model used for processing graph data, which can effectively capture structural and feature information in graph data. Anomaly detection refers to identifying unexpected data from a massive amount of data. Traditional anomaly detection methods usually do not consider the relationships between data when detecting graph data, while models that use GNN for anomaly detection can learn from graph structures and features, thereby improving the accuracy and robustness of anomaly detection. This paper reviews the application of GNN in anomaly detection from three aspects. Firstly, the basic framework of GNN is introduced. Secondly, the latest research progress of GNN in static graph anomaly detection, dynamic graph anomaly detection, and time series data anomaly detection is discussed separately. Finally, an in-depth analysis is conducted on the future research directions in this field. Reference | Related Articles | Metrics Select Review of Research on Application of Transformer in Domain Adaptation CHEN Jianwei, YU Lu, HAN Changzhi, LI Lin Computer Engineering and Applications    2024, 60 (13): 66-80.   DOI: 10.3778/j.issn.1002-8331.2310-0290 Abstract （59）      PDF（pc） （5728KB）（75）       Knowledge map    Save Domain adaptation, the important branch of transfer learning, aims to solve the problem that the performance of traditional machine learning algorithms drops sharply when the training and test samples obey different data distributions. Transformer is a deep learning framework based on a self-attention mechanism, which has strong global feature extraction ability and modeling ability. In recent years, the combination of Transformer and domain adaptation has also become a research hotspot. Although many relative methods have been published, the review of Transformer application in domain adaptation has not been reported. In order to fill the gap in this field and provide reference for relevant research, this paper summarizes and analyzes some typical domain adaptation methods based on Transformer in recent years. This paper summarizes the concepts related to domain adaptation and the basic structure of the Transformer, sorts out various domain adaptation methods based on Transformer from four applications, i.e. image classification, image semantic segmentation, object detection and medical image analysis and compares the domain adaptation methods in image classification. Finally, the challenges of the current domain adaptation Transformer model are summarized, and the feasible research directions in the future are discussed. Reference | Related Articles | Metrics Select Review of Connected Autonomous Vehicle Cooperative Control at On-Ramp Merging Areas LI Chun, WU Zhizhou, ZENG Guang, ZHAO Xin, YANG Zhidan Computer Engineering and Applications    2024, 60 (12): 1-17.   DOI: 10.3778/j.issn.1002-8331.2310-0310 Abstract （134）      PDF（pc） （5963KB）（229）       Knowledge map    Save The area where vehicles conduct interchanges is designated as the on-ramp merging area. The traffic efficiency in the ramp merging area drastically decreases if the mainline and ramp traffic flow density reaches saturation. As a current research hotspot in transportation, intelligent network technology, relying on the high-precision motion control and high-efficiency communication of connected-automated vehicle (CAV), can significantly improve the traffic efficiency in the merging area. The fusion strategies used by CAV are assessed in this research utilizing three different control paradigms: feedback control, optimal control, and reinforcement learning. The shortcomings of the three methods in this scenario are summarized, and specific improvement measures are given by reviewing existing research. Also, it offers a thorough summary of the most recent developments and trends in this particular scientific field. Reference | Related Articles | Metrics Select Survey of Deep Learning Based Approaches for Gaze Estimation WEN Mingqi, REN Luqian, CHEN Zhenqin, YANG Zhuo, ZHAN Yinwei Computer Engineering and Applications    2024, 60 (12): 18-33.   DOI: 10.3778/j.issn.1002-8331.2309-0497 Abstract （117）      PDF（pc） （6991KB）（196）       Knowledge map    Save Gaze estimation is a technique for predicting the gaze position or gaze direction of the human eye and plays an important role in human-computer interaction and computer vision applications. The recent development of deep learning has revolutionized many computer vision tasks, and using deep learning for appearance-based gaze estimation has also become a hot topic. Focusing on the training process of the deep learning model, this paper analyzes state-of-the-art gaze estimation methods from four perspectives: gaze data preprocessing, gaze feature extraction, gaze learning strategies, and deep gaze model structures. In addition, the mainstream public datasets are summarized, and the performance evaluation and analysis of 2D and 3D gaze estimation methods are carried out on several popular datasets. Finally, the challenges faced by the existing gaze estimation methods are discussed, and the future development directions are prospected. Reference | Related Articles | Metrics Select Advancements in Embedded Static Knowledge Graph Completion Research WU Yujie, XI Xuefeng, CUI Zhiming Computer Engineering and Applications    2024, 60 (12): 34-47.   DOI: 10.3778/j.issn.1002-8331.2310-0221 Abstract （54）      PDF（pc） （5038KB）（65）       Knowledge map    Save Knowledge graphs are widely used and semantically rich data representations, which is increasingly becoming a crucial technology in the field of knowledge engineering. However, real-world knowledge graphs often suffer from incompleteness and ambiguity, hindering their application performance. Knowledge graph completion techniques aim to enrich the content of knowledge graphs by predicting missing entities or relations, has been a hot research topic in recent years. In particular, embedding-based approaches have made remarkable progress in knowledge graph completion tasks. It reviews recent embedding-based static knowledge graph completion methods, categorizing them based on approaches such as translation-based models, tensor factorization, neural network models, and pre-trained language models. These methods achieve an improved semantic representation and inferential capabilities by embedding entity relations into continuous vector spaces. At the same time, it has potential advantages in capturing complex relationships between entities and utilizing graph structural information. Reference | Related Articles | Metrics Select Review of Research on Facial Landmark Detection ZHANG Xiaohang, TIAN Qichuan, LIAN Lu, TAN Run Computer Engineering and Applications    2024, 60 (12): 48-60.   DOI: 10.3778/j.issn.1002-8331.2311-0387 Abstract （66）      PDF（pc） （5543KB）（82）       Knowledge map    Save With the rapid development of computer vision and other technologies, the rapid rise of human-computer interaction, medical assistance, security monitoring and other fields, facial landmark detection as one of the important tasks of concern, it can locate and detect facial landmark in images or videos, with high practical value. By combing and analyzing the current research status of face key point detection methods, this paper divides them into traditional facial landmark detection methods and deep learning-based facial landmark detection methods.It compares and analyzes the principles, advantages and disadvantages of various methods, introduces common data sets and evaluation indicators, and comprehensively evaluates the performance of key methods on different data sets. The application field of facial landmark detection is summarized, and its future development direction is forecasted. Reference | Related Articles | Metrics Select Overview of Knowledge Graph Completion Methods ZHANG Wenhao, XU Zhenshun, LIU Na, WANG Zhenbiao, TANG Zengjin, WANG Zheng’an Computer Engineering and Applications    2024, 60 (12): 61-73.   DOI: 10.3778/j.issn.1002-8331.2311-0029 Abstract （73）      PDF（pc） （4635KB）（89）       Knowledge map    Save Knowledge graph is a semantic network used to describe various entities and concepts that exist in the world, as well as their relationships. In recent years, it has been widely used in fields such as intelligent question answering, intelligent recommendation, and information retrieval. At present, most knowledge graphs are incomplete, therefore, knowledge graph completion has become an important task. Firstly, based on the different methods of model construction, knowledge graph completion models are divided into three categories: traditional knowledge graph completion models, knowledge graph completion models based on neural networks, and knowledge graph completion models based on meta learning. The classification of these three knowledge graph completion models is introduced. Then, the dataset and evaluation indicators used in the knowledge graph completion method are summarized, and detailed comparative analysis is conducted on various models from the perspectives of their advantages and disadvantages. Finally, the knowledge graph completion is summarized and summarized, and future research directions are prospected. Reference | Related Articles | Metrics Select Survey on Identity Authentication in Metaverse Environment DENG Miaolei, ZHAI Haonan, MA Mimi, ZUO Zhibin Computer Engineering and Applications    2024, 60 (11): 1-16.   DOI: 10.3778/j.issn.1002-8331.2309-0280 Abstract （123）      PDF（pc） （5589KB）（133）       Knowledge map    Save The identity authentication schemes in the traditional network environment are designed according to the characteristics of their own environments, which are difficult to meet the design requirements in the metaverse environment. In order to solve the problem of identity authentication in the metaverse environment, the characteristics, attributes, application scenarios and main attack threats faced by the metaverse are studied. The identity privacy and security risks related to users, avatars, platforms, wearable devices and communications in the metaverse are discussed. The work related to identity authentication in the metaverse environment is sorted out and categorized into user-avatar, user-platform and user-wearable device authentication. Finally, the research directions of identity authentication mechanism in the metaverse environment are discussed. Reference | Related Articles | Metrics Select Review of Research on Structure and Autonomous Control of Water Surface Garbage Cleaning Robots PAN Zijun, WANG Jianhua, ZHENG Xiang, TIAN Yu, TIAN Yi’ning, ZHANG Mengdi, WANG Haozhu, CHE Wenbo Computer Engineering and Applications    2024, 60 (11): 17-31.   DOI: 10.3778/j.issn.1002-8331.2310-0352 Abstract （101）      PDF（pc） （5839KB）（97）       Knowledge map    Save With the amount of garbage on water surface ever increasing, traditional cleaning methods are no longer able to meet the requirements for safety, efficiency, as well as intelligence in applications. Hence, water surface garbage cleaning robots have emerged and become the focus of development in this field. After more than 20 years of development, it has entered the stage of practical application from research with various structure types designed and the realization of fully autonomous operations from garbage detection to collection. A large amount of literature has been accumulated, but there is still no comprehensive review. Therefore, this paper aims to provide an extensive review of development of water surface garbage cleaning robots in literature. The structures of the water surface garbage cleaning robots are classified with the characteristics of each type summarized. The typical framework of an autonomous water surface garbage cleaning robot is analyzed, and its three key technologies of perception, decision-making, and control are highlighted and discussed. Finally, the paper explores the trends of development for technologies in water surface garbage cleaning robots, which are centered on the challenges presented by current applications, to provide a reference for further research and application of water surface garbage cleaning robots. Reference | Related Articles | Metrics Select Review of Medical Image Segmentation Algorithms Based on U-Net Variants CUI Ke, TIAN Qichuan, LIAN Lu Computer Engineering and Applications    2024, 60 (11): 32-49.   DOI: 10.3778/j.issn.1002-8331.2310-0335 Abstract （108）      PDF（pc） （6802KB）（193）       Knowledge map    Save The simple and efficient network structure of U-Net is widely used in medical image segmentation, and many scholars have made various researches on the U-Net structure. This paper elucidates in the following: firstly, the paper summarizes the key challenges of the U-Net network in the field of medical image segmentation; next, it elaborates the formats and characteristics of medical image datasets that are commonly used in the U-Net network; then, it summarizes the six improvement mechanism of U-Net：skip connection mechanism, generative adversarial network, residual connection mechanism, 3D-UNet, Transformer mechanism, and dense connecting mechanism. Finally, the paper discusses the relationship between these improvement mechanisms and commonly used medical data formats, and points out the ideas and directions for future improvement, so as to stimulate the unlimited potential of U-Net in medical image segmentation. Reference | Related Articles | Metrics Select Text Classification：Comprehensive Review of Prompt Learning Methods GU Xunxun, LIU Jianping, XING Jialu, REN Haiyu Computer Engineering and Applications    2024, 60 (11): 50-61.   DOI: 10.3778/j.issn.1002-8331.2310-0049 Abstract （61）      PDF（pc） （4323KB）（62）       Knowledge map    Save Text classification is a basic task in natural language processing, which has important applications in sentiment analysis, news classification and other fields. Compared with traditional machine learning and deep learning models, prompt learning can construct prompts for text classification in the case of insufficient data. In recent years, the emergence of GPT-3 has promoted the development of cue learning methods, and has made significant progress in the field of text classification. Firstly, this paper briefly combs the process of previous text classification methods and analyzes their existing problems and shortcomings. Secondly, it expounds the development process of cue learning and the method of constructing cue templates, and introduces and summarizes the research and results of cue learning methods for text classification. Finally, the development trend and difficulties to be further studied in the field of text classification are summarized and prospected. Reference | Related Articles | Metrics Select Survey of Specific Speech Recognition Algorithms for Dysarthria SONG Wei, ZHANG Yanghao Computer Engineering and Applications    2024, 60 (11): 62-74.   DOI: 10.3778/j.issn.1002-8331.2309-0154 Abstract （49）      PDF（pc） （5779KB）（44）       Knowledge map    Save Articulation disorder, as a medical difficulty, currently mainstream speech recognition technologies are not well adapted to the needs of this field. At the same time, speech recognition technology for dysarthria utilizes a combination of pre training and personalized training to further improve algorithm performance and reduce recognition word error rate through data-driven methods. However, currently, speech recognition technology for dysarthria still has a certain distance from practical commercial use, and its development is limited by data scale and technology. So far, there have been no comprehensive articles on speech recognition for dysarthria. It is urgent to compare and analyze the construction methods and advanced technologies of various datasets in this field, in order to facilitate researchers entering the field to quickly acquire knowledge in this field. This paper conducts a survey on existing datasets, mainstream algorithms, and evaluation methods, and summarizes the scale, form, and characteristics of mainstream speech impairment datasets at home and abroad. It analyzes the mainstream algorithms for speech recognition with dysarthria, and provides the performance and characteristics of different algorithms. Finally, the performance evaluation indicators of the algorithm model based on the severity level of patients with dysarthria are studied, and future research directions are discussed, in order to provide help for the researchers engaged in speech recognition with dysarthria and assist in the rapid development of this field. Reference | Related Articles | Metrics Select Process of Weakly Supervised Salient Object Detection YU Junwei, GUO Yuansen, ZHANG Zihao, MU Yashuang Computer Engineering and Applications    2024, 60 (10): 1-15.   DOI: 10.3778/j.issn.1002-8331.2308-0206 Abstract （169）      PDF（pc） （6029KB）（246）       Knowledge map    Save Salient object detection aims to accurately detect and locate the most attention-grabbing objects or regions in images or videos, facilitating better object recognition and scene analysis. Despite the effectiveness of fully supervised saliency detection methods, acquiring large pixel-level annotated datasets is challenging and costly. Weakly supervised detection methods utilize relatively easy-to-obtain image-level labels or noisy weak labels to train models, demonstrating good performance in practical applications. This paper comprehensively compares the mainstream methods and application scenarios of fully supervised and weakly supervised saliency detection methods, and then analyzes the data annotation methods using weak labels and their impact on salient object detection. The latest research progress in salient object detection under weakly supervised conditions is reviewed, and the performance of various weakly supervised methods is compared on several public datasets. Finally, the potential applications of weakly supervised saliency detection methods in special fields such as agriculture, medicine and military are discussed, highlighting the existing challenges and future development trends in this research area. Reference | Related Articles | Metrics Select Review of Research on Local Path Planning for Intelligent Construction Robots YU Junqi, CHEN Yisheng, FENG Chunyong, SU Yucong, GUO Jugang Computer Engineering and Applications    2024, 60 (10): 16-29.   DOI: 10.3778/j.issn.1002-8331.2309-0203 Abstract （103）      PDF（pc） （5010KB）（138）       Knowledge map    Save Intelligent construction robots, as the key equipment to implement the intelligent transformation of the construction industry, represent the level of intelligent construction with their autonomous construction capability, and local path planning is the key technology for efficient robot construction. Aiming at the environmental characteristics of the construction site, the difficulties of local path planning for intelligent construction robots are explored. Firstly, classical local path planning algorithms such as artificial potential field, time elasticity band, Bug, dynamic window, etc. are analyzed. Secondly, local path planning algorithms of artificial intelligence such as reinforcement learning, deep reinforcement learning, fuzzy control, swarm intelligence, etc. are summarized. Finally, the limitations of various methods on the application of intelligent construction robots in the construction site environment are analyzed, and the development trend of local path planning algorithms of intelligent construction robots is discussed, aiming to provide certain ideas and suggestions for the research on local path planning of intelligent construction robots in complex construction sites, improve the autonomous construction ability of robots, and promote the development of intelligent construction technology. Reference | Related Articles | Metrics Select Survey of  Vision Transformer in Fine-Grained Image Classification SUN Lulu, LIU Jianping, WANG Jian, XING Jialu, ZHANG Yue, WANG Chenyang Computer Engineering and Applications    2024, 60 (10): 30-46.   DOI: 10.3778/j.issn.1002-8331.2310-0395 Abstract （108）      PDF（pc） （11071KB）（135）       Knowledge map    Save Fine-grained image classification (FGIC) has always been an important problem in computer vision. Compared to traditional image classification tasks, FGIC faces the challenge of extremely similar inter-class objects, which further increases the difficulty of the task. With the development of deep learning, Vision Transformer (ViT) models have become popular in the field of vision and have been introduced into FGIC tasks. This paper introduces the challenges faced by FGIC tasks, provides an overview of the ViT model, and analyzes its characteristics. The comprehensive review is primarily based on the model structure and covers FGIC algorithms based on ViT. It includes feature extraction, feature relation modeling, feature attention, and feature enhancement as the main aspects. Each algorithm is summarized, and its advantages and disadvantages are analyzed. Following that, a comparison of the performance of different ViT models on the same public dataset is conducted to validate their effectiveness in the FGIC tasks. Furthermore, the limitations of current research are pointed out, and future research directions are proposed to further explore the potential of ViT in FGIC. Reference | Related Articles | Metrics Select Review of Research on Neural Network Combined with Attention Mechanism in Recommendation System GAO Guangshang Computer Engineering and Applications    2024, 60 (10): 47-60.   DOI: 10.3778/j.issn.1002-8331.2308-0014 Abstract （74）      PDF（pc） （4771KB）（122）       Knowledge map    Save Explore how neural networks combine attention mechanisms and their variants to better learn complex and implicit relationships between users and items, thereby improving the accuracy and personalization of recommendations. Starting from six typical types of neural networks: multi-layer perceptron, convolutional neural network, recurrent neural network, autoencoder, graph neural network, and backpropagation neural network, this paper studies the process of combining them with the attention mechanism for recommendation. Specifically, the advantages and disadvantages are analyzed based on typical application scenarios such as click-through rate prediction, tag recommendation, and review rating prediction. By combining neural networks with attention mechanisms, the model can focus on key information in the input data, reduce attention to secondary information, and even directly filter out irrelevant information. Existing recommendation models that combine attention mechanisms with neural networks, to a large extent, can meet the needs of common recommendation tasks. However, this type of model still faces some challenges in complex recommendation scenarios such as cross-domain recommendation, deep reinforcement learning recommendation, and multi-modal recommendation. For example, cross-domain recommendation requires the model with the ability of transfer learning, and reinforcement learning recommendation requires long-term reward modeling. Reference | Related Articles | Metrics Select Survey of Physical Adversarial Attacks Against Object Detection Models CAI Wei, DI Xingyu, JIANG Xinhao, WANG Xin, GAO Weijie Computer Engineering and Applications    2024, 60 (10): 61-75.   DOI: 10.3778/j.issn.1002-8331.2310-0362 Abstract （63）      PDF（pc） （7385KB）（99）       Knowledge map    Save Deep learning models are highly susceptible to adversarial samples, and even minuscule image perturbations that are not perceptible to the naked eye can disable well-trained deep learning models. Recent research indicates that these perturbations can exist in the physical world. This paper provides insight into physical adversarial attacks on deep learning object detection models, clarifying the concept of physical adversarial attack and outlining the general process of such attacks on object detection. According to the different attack tasks, a series of physical adversarial attack methods against object detection networks in recent years are reviewed from vehicle detection and pedestrian detection. Other attacks against target detection models, other attack tasks and other attack methods are briefly introduced. The current challenges of physical adversarial attack are discussed, the limitations of adversarial training are leaded out, and future development directions and application prospect are suggested. Reference | Related Articles | Metrics page Page 1 of 9 Total 342 records First page Prev page Next page Last page