Jun Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Jun Liu has a total of 68 co-authors. Here are the top co-authors by number of shared publications:
- Mingwei Lin
Fujian Normal University
9 shared publications Guangxia Xu
6 shared publications- Mingyue Xie
School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu, China
6 shared publications Shuyu Chen
School of Big Data and Software Engineering, Chongqing University, Chongqing, China
5 shared publications- Guangxia Xu
Guangzhou University
3 shared publications - Zeshui Xu
Sichuan University
3 shared publications
Latest Publications
- 23 February 2026
- Published by Springer Nature
- p. 332-347
MeDRNet: A Knowledge-Augmented Multi-model Framework for Robust Medical Language Understanding
- 23 February 2026
- Published by Springer Nature
- p. 332-347
MeDRNet: A Knowledge-Augmented Multi-model Framework for Robust Medical Language Understanding
- 1 January 2026
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 87 (1) , 1-10
A Comprehensive Literature Review on YOLO-Based Small Object Detection: Methods, Challenges, and Future Trends
Small object detection has been a focus of attention since the emergence of deep learning-based object detection. Although classical object detection frameworks have made significant contributions to the development of object detection, there are still many issues to be resolved in detecting small objects due to the inherent complexity and diversity of real-world visual scenes. In particular, the YOLO (You Only Look Once) series of detection models, renowned for their real-time performance, have undergone numerous adaptations aimed at improving the detection of small targets. In this survey, we summarize the state-of-the-art YOLO-based small object detection methods. This review presents a systematic categorization of YOLO-based approaches for small-object detection, organized into four methodological avenues, namely attention-based feature enhancement, detection-head optimization, loss function, and multi-scale feature fusion strategies. We then examine the principal challenges addressed by each category. Finally, we analyze the performance of these methods on public benchmarks and, by comparing current approaches, identify limitations and outline directions for future research.
- 1 January 2026
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 87 (1) , 1-10
A Comprehensive Literature Review on YOLO-Based Small Object Detection: Methods, Challenges, and Future Trends
Small object detection has been a focus of attention since the emergence of deep learning-based object detection. Although classical object detection frameworks have made significant contributions to the development of object detection, there are still many issues to be resolved in detecting small objects due to the inherent complexity and diversity of real-world visual scenes. In particular, the YOLO (You Only Look Once) series of detection models, renowned for their real-time performance, have undergone numerous adaptations aimed at improving the detection of small targets. In this survey, we summarize the state-of-the-art YOLO-based small object detection methods. This review presents a systematic categorization of YOLO-based approaches for small-object detection, organized into four methodological avenues, namely attention-based feature enhancement, detection-head optimization, loss function, and multi-scale feature fusion strategies. We then examine the principal challenges addressed by each category. Finally, we analyze the performance of these methods on public benchmarks and, by comparing current approaches, identify limitations and outline directions for future research.
- 6 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 13 (2) , 1513-1528
A Scalable Multichannel Sentiment Analysis Model With Enhanced Semantic Understanding and Redundancy Reduction
The prevailing techniques in textual sentiment analysis include convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM) networks, and attention mechanisms. However, the existing sentiment analysis methods typically focus on relationships among neighboring words, neglecting long-range dependencies. Moreover, they fail to reduce redundant information in semantic features, ultimately affecting the effectiveness of sentiment analysis. To tackle these challenges, this article introduces a novel text sentiment analysis model consisting of three key modules: word embedding, a scalable multichannel sentiment feature extractor, and a sentiment classifier. The scalable multichannel sentiment feature extractor employs a scalable N-gram dilated CNN (SN-DCNN) to extract multiscale high-level semantic by capturing both neighboring and nonneighboring word associations. To further enhance the model’s ability to capture semantic features, a BiLSTM network is introduced to capture long-range dependencies within sentences. Subsequently, a structured attention (SA) mechanism is incorporated after the BiLSTM to capture multiscale semantic information, thereby enabling the mapping of individual long-range dependencies to distinct semantic levels. In addition, a semantics-driven channel attention mechanism (SDCAM) is introduced to assign weights to multichannel structures, aiming to reduce the redundancy in the semantic features. Furthermore, by integrating multihead self-attention (MHSA), the model can effectively concentrate on key sentiment-related features, resulting in a comprehensive enhancement of its semantic representation capability. The sentiment classifier utilizes an adaptive loss function based on cross-entropy to correct the probability distribution of sentiment classes. This function enables the model to dynamically adjust its learning capacity for each class, leading to significant performance improvements, especially in datasets with imbalanced class distributions. Experimental results show that the proposed model improves accuracy by 0.94%, 0.96%, 0.36%, and 0.86% on the NLPCC2017-ECGC, ChnSentiCorp, COLD, and IMDB datasets, respectively, when using word-level embeddings. With character-level embeddings, the model improves accuracy by 0.76% and 1.23% on the ChnSentiCorp and COLD datasets, respectively, compared to the state-of-the-art methods.
- 6 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 13 (2) , 1513-1528
A Scalable Multichannel Sentiment Analysis Model With Enhanced Semantic Understanding and Redundancy Reduction
The prevailing techniques in textual sentiment analysis include convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM) networks, and attention mechanisms. However, the existing sentiment analysis methods typically focus on relationships among neighboring words, neglecting long-range dependencies. Moreover, they fail to reduce redundant information in semantic features, ultimately affecting the effectiveness of sentiment analysis. To tackle these challenges, this article introduces a novel text sentiment analysis model consisting of three key modules: word embedding, a scalable multichannel sentiment feature extractor, and a sentiment classifier. The scalable multichannel sentiment feature extractor employs a scalable N-gram dilated CNN (SN-DCNN) to extract multiscale high-level semantic by capturing both neighboring and nonneighboring word associations. To further enhance the model’s ability to capture semantic features, a BiLSTM network is introduced to capture long-range dependencies within sentences. Subsequently, a structured attention (SA) mechanism is incorporated after the BiLSTM to capture multiscale semantic information, thereby enabling the mapping of individual long-range dependencies to distinct semantic levels. In addition, a semantics-driven channel attention mechanism (SDCAM) is introduced to assign weights to multichannel structures, aiming to reduce the redundancy in the semantic features. Furthermore, by integrating multihead self-attention (MHSA), the model can effectively concentrate on key sentiment-related features, resulting in a comprehensive enhancement of its semantic representation capability. The sentiment classifier utilizes an adaptive loss function based on cross-entropy to correct the probability distribution of sentiment classes. This function enables the model to dynamically adjust its learning capacity for each class, leading to significant performance improvements, especially in datasets with imbalanced class distributions. Experimental results show that the proposed model improves accuracy by 0.94%, 0.96%, 0.36%, and 0.86% on the NLPCC2017-ECGC, ChnSentiCorp, COLD, and IMDB datasets, respectively, when using word-level embeddings. With character-level embeddings, the model improves accuracy by 0.76% and 1.23% on the ChnSentiCorp and COLD datasets, respectively, compared to the state-of-the-art methods.
- Open Access
TFM Dataset: A Novel Multi-task Dataset and Integrated Pipeline for Automated Tear Film Break-Up Segmentation
Tear film break-up (TFBU) analysis is critical for diagnosing dry eye syndrome, but automated TFBU segmentation remains challenging due to the lack of annotated datasets and integrated solutions. This paper introduces the Tear Film Multi-task (TFM) Dataset, the first comprehensive dataset for multi-task tear film analysis, comprising 15 high-resolution videos (totaling 6,247 frames) annotated with three vision tasks: frame-level classification ('clear', 'closed', 'broken', 'blur'), Placido Ring detection, and pixel-wise TFBU area segmentation. Leveraging this dataset, we first propose TF-Net, a novel and efficient baseline segmentation model. TF-Net incorporates a MobileOne-mini backbone with re-parameterization techniques and an enhanced feature pyramid network to achieve a favorable balance between accuracy and computational efficiency for real-time clinical applications. We further establish benchmark performance on the TFM segmentation subset by comparing TF-Net against several state-of-the-art medical image segmentation models. Furthermore, we design TF-Collab, a novel integrated real-time pipeline that synergistically leverages models trained on all three tasks of the TFM dataset. By sequentially orchestrating frame classification for BUT determination, pupil region localization for input standardization, and TFBU segmentation, TF-Collab fully automates the analysis. Experimental results demonstrate the effectiveness of the proposed TF-Net and TF-Collab, providing a foundation for future research in ocular surface diagnostics. Our code and the TFM datasets are available at https://github.com/glory-wan/TF-Net
Open AccessTFM Dataset: A Novel Multi-task Dataset and Integrated Pipeline for Automated Tear Film Break-Up Segmentation
Tear film break-up (TFBU) analysis is critical for diagnosing dry eye syndrome, but automated TFBU segmentation remains challenging due to the lack of annotated datasets and integrated solutions. This paper introduces the Tear Film Multi-task (TFM) Dataset, the first comprehensive dataset for multi-task tear film analysis, comprising 15 high-resolution videos (totaling 6,247 frames) annotated with three vision tasks: frame-level classification ('clear', 'closed', 'broken', 'blur'), Placido Ring detection, and pixel-wise TFBU area segmentation. Leveraging this dataset, we first propose TF-Net, a novel and efficient baseline segmentation model. TF-Net incorporates a MobileOne-mini backbone with re-parameterization techniques and an enhanced feature pyramid network to achieve a favorable balance between accuracy and computational efficiency for real-time clinical applications. We further establish benchmark performance on the TFM segmentation subset by comparing TF-Net against several state-of-the-art medical image segmentation models. Furthermore, we design TF-Collab, a novel integrated real-time pipeline that synergistically leverages models trained on all three tasks of the TFM dataset. By sequentially orchestrating frame classification for BUT determination, pupil region localization for input standardization, and TFBU segmentation, TF-Collab fully automates the analysis. Experimental results demonstrate the effectiveness of the proposed TF-Net and TF-Collab, providing a foundation for future research in ocular surface diagnostics. Our code and the TFM datasets are available at https://github.com/glory-wan/TF-Net
- 11 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 26 (12) , 21839-21856
BPSAC: A Blockchain and PUF-Enabled Security Architecture With C-IDS for Maritime Intelligent Transportation Systems
Blockchain-enabled security mechanisms play a crucial role in Maritime Intelligent Transportation Systems (MITSs). However, the growing dependence on network communication and data exchange raises serious privacy concerns. While numerous authentication schemes exist for MITSs to prevent unauthorized access, most exhibit limited scalability, high computational costs, and vulnerability to physical attacks, hindering their practical implementation. Still, the transmission of authentication information relies on open networks, exposing the system to diverse network attacks. To cope with these issues, a blockchain-enabled security architecture with Physically Unclonable Functions (PUFs) and consensus-based intrusion detection system (BPSAC) is proposed to protect privacy and ensure resilience against various network attacks. The architecture comprises a security authentication mechanism and a novel consensus-based intrusion detection system (C-IDS). Specifically, the security authentication mechanism mitigates performance overhead and privacy concerns by introducing PUFs. Subsequently, C-IDS is designed for intrusion detection without sacrificing the security of the model inference process by introducing a Differential Channel and Spatial Attention (DCSA) module, and an enhanced Practical Byzantine Fault Tolerance (PBFT) consensus algorithm named MI-PBFT. Experimental results demonstrate that the proposed security authentication mechanism reduces the computational overhead by at least 20% and the communication overhead by 5% compared to recent works. The proposed detection model achieves an attack detection accuracy of over 99% on the UNSW-NB15, KDDCup99, and NSL-KDD datasets. Furthermore, as the number of consensus nodes increases, MI-PBFT consensus algorithm reduces the transaction confirmation delay by 32% and improves the transactions per second by 80% compared to PBFT.
- 11 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 26 (12) , 21839-21856
BPSAC: A Blockchain and PUF-Enabled Security Architecture With C-IDS for Maritime Intelligent Transportation Systems
Blockchain-enabled security mechanisms play a crucial role in Maritime Intelligent Transportation Systems (MITSs). However, the growing dependence on network communication and data exchange raises serious privacy concerns. While numerous authentication schemes exist for MITSs to prevent unauthorized access, most exhibit limited scalability, high computational costs, and vulnerability to physical attacks, hindering their practical implementation. Still, the transmission of authentication information relies on open networks, exposing the system to diverse network attacks. To cope with these issues, a blockchain-enabled security architecture with Physically Unclonable Functions (PUFs) and consensus-based intrusion detection system (BPSAC) is proposed to protect privacy and ensure resilience against various network attacks. The architecture comprises a security authentication mechanism and a novel consensus-based intrusion detection system (C-IDS). Specifically, the security authentication mechanism mitigates performance overhead and privacy concerns by introducing PUFs. Subsequently, C-IDS is designed for intrusion detection without sacrificing the security of the model inference process by introducing a Differential Channel and Spatial Attention (DCSA) module, and an enhanced Practical Byzantine Fault Tolerance (PBFT) consensus algorithm named MI-PBFT. Experimental results demonstrate that the proposed security authentication mechanism reduces the computational overhead by at least 20% and the communication overhead by 5% compared to recent works. The proposed detection model achieves an attack detection accuracy of over 99% on the UNSW-NB15, KDDCup99, and NSL-KDD datasets. Furthermore, as the number of consensus nodes increases, MI-PBFT consensus algorithm reduces the transaction confirmation delay by 32% and improves the transactions per second by 80% compared to PBFT.
- 2 September 2025
- Published by Elsevier in Computer Networks
- Vol. 272
Comprehensive survey of blockchain consensus mechanisms: Analysis, applications, and future trends
- 2 September 2025
- Published by Elsevier in Computer Networks
- Vol. 272
Comprehensive survey of blockchain consensus mechanisms: Analysis, applications, and future trends
- 20 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Fuzzy Systems
- Vol. 33 (9) , 3061-3075
Anisotropic-Gaussian-Kernel-Based Fuzzy Clustering Algorithm for Feature Selection
Soft clustering algorithms based on fuzzy C-means (FCM) have been extensively applied to complex data analysis. However, existing FCM variants still encounter key limitations: a large number of iterations due to slow convergence on highdimensional data, equal weights of all samples which increases sensitivity to noise, and strong dependence on empirically chosen fuzzy parameters, often resulting in suboptimal solutions. In order to address these challenges, in this work, we propose an adaptive FCM clustering algorithm based on anisotropic Gaussian kernel function (AGK-FCM for short), which facilitates the process of feature selection for multidimensional data with fewer iterations after feature reduction based on updating the kernel width vector. In order to enhance the classification accuracy, we assign adaptive weights to each sample and adjust these weights to mitigate the impact of outliers on classification accuracy. Unlike traditional fuzzy clustering algorithms, we employ the PSO-TVAC algorithm to determine the global optimal parameters, thus reducing the computational resources and enhancing the algorithm's efficiency. The experimental results based on 16 publicly available datasets validate that the proposed algorithm achieves higher classification accuracy with minimal number of iterations.
- 20 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Fuzzy Systems
- Vol. 33 (9) , 3061-3075
Anisotropic-Gaussian-Kernel-Based Fuzzy Clustering Algorithm for Feature Selection
Soft clustering algorithms based on fuzzy C-means (FCM) have been extensively applied to complex data analysis. However, existing FCM variants still encounter key limitations: a large number of iterations due to slow convergence on highdimensional data, equal weights of all samples which increases sensitivity to noise, and strong dependence on empirically chosen fuzzy parameters, often resulting in suboptimal solutions. In order to address these challenges, in this work, we propose an adaptive FCM clustering algorithm based on anisotropic Gaussian kernel function (AGK-FCM for short), which facilitates the process of feature selection for multidimensional data with fewer iterations after feature reduction based on updating the kernel width vector. In order to enhance the classification accuracy, we assign adaptive weights to each sample and adjust these weights to mitigate the impact of outliers on classification accuracy. Unlike traditional fuzzy clustering algorithms, we employ the PSO-TVAC algorithm to determine the global optimal parameters, thus reducing the computational resources and enhancing the algorithm's efficiency. The experimental results based on 16 publicly available datasets validate that the proposed algorithm achieves higher classification accuracy with minimal number of iterations.
- 1 May 2025
- Published by Springer Nature
- p. 3-18
An Improved DPoS Consensus Mechanism in Blockchain Based on PDHLTS
- 1 May 2025
- Published by Springer Nature
- p. 3-18
An Improved DPoS Consensus Mechanism in Blockchain Based on PDHLTS
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 22 (5) , 5335-5351
Traceability and Identity Protection in Smart Agricultural IoT System Framework Based on Blockchains
Blockchain-based IoT applications in agriculture have drawn extensive attention in recent years, allowing the implementation of smart agriculture solutions. By transmitting collected relevant data to a control center through the blockchain, corresponding regulation can be realized in agricultural production management systems. However, existing efforts for directly adopting the technique to data transmission are obstructed by several issues. The traceability of agricultural data stored in the blockchain leads to the exposure of the identity of the data collecting devices. And the tracing difficulty of completely invisible data for identity protection also exists in the smart agricultural system. To tackle these limitations, we propose a novel blockchain-based smart agricultural IoT system framework for regulating the agricultural production environment through trusted data. First, the elliptic curve integrated encryption scheme (ECIES) and the group signature scheme are integrated to guarantee the traceability and identity protection of the data and equipment, respectively. Then, to enhance the security of session key transmission in the ECIES scheme, we further design an on-chain-off-chain key agreement protocol (ECIES-OOKA). In addition, we propose a novel group manager selection method based on probabilistic linguistic term sets (PLTSs) for the group signature implementation. Finally, a practical example is provided to demonstrate the group manager selection process and verify the feasibility of the proposed method. The security and performance analysis for the system framework are also presented.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 22 (5) , 5335-5351
Traceability and Identity Protection in Smart Agricultural IoT System Framework Based on Blockchains
Blockchain-based IoT applications in agriculture have drawn extensive attention in recent years, allowing the implementation of smart agriculture solutions. By transmitting collected relevant data to a control center through the blockchain, corresponding regulation can be realized in agricultural production management systems. However, existing efforts for directly adopting the technique to data transmission are obstructed by several issues. The traceability of agricultural data stored in the blockchain leads to the exposure of the identity of the data collecting devices. And the tracing difficulty of completely invisible data for identity protection also exists in the smart agricultural system. To tackle these limitations, we propose a novel blockchain-based smart agricultural IoT system framework for regulating the agricultural production environment through trusted data. First, the elliptic curve integrated encryption scheme (ECIES) and the group signature scheme are integrated to guarantee the traceability and identity protection of the data and equipment, respectively. Then, to enhance the security of session key transmission in the ECIES scheme, we further design an on-chain-off-chain key agreement protocol (ECIES-OOKA). In addition, we propose a novel group manager selection method based on probabilistic linguistic term sets (PLTSs) for the group signature implementation. Finally, a practical example is provided to demonstrate the group manager selection process and verify the feasibility of the proposed method. The security and performance analysis for the system framework are also presented.