Liming Sun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Liming Sun has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Jie Li
Northeastern University
3 shared publications - Bo Yi
Northeastern University
2 shared publications - Ruiyun Yu
Northeastern University
2 shared publications - Yuting Wang
Software College, Northeastern University, Shenyang 110819, China
2 shared publications - Bang An
Software College, Northeastern University, Shenyang 110819, China
1 shared publication - Bingyang Guo
Software College, Northeastern University, Shenyang 110819, China
1 shared publication
Latest Publications
- 11 November 2026
- Published by IOP Publishing in Engineering Research Express
CARDNet: A Lightweight Regeneration-Aware Network for ElectricVehicle Battery State-of-Health Estimation from Weekly Telematics Data
Accurate state-of-health (SOH) estimation is essential for safe and economical electric vehicle (EV) operation, yet real-world fleet telematics typically provides short, weekly aggregated sequences with pronounced capacity regeneration and substantial cross-vehicle heterogeneity, challenging deep models developed for dense laboratory data. This paper proposes CARDNet, a lightweight network that isolates high-frequency input fluctuations, where regeneration dynamics are concentrated, through a parameter-free moving-average decomposition. The extracted component is encoded by a compact convolutional branch and complemented by a global feature-interaction path that captures cross-feature nonlinearities. We further provide an explicit operational definition of capacity regeneration, under which the studied fleet exhibits 117 regeneration episodes that transiently reverse capacity equal to 48.8% of the net decay. Experiments on 20 real EVs monitored for over two years use a unified one-week-ahead prediction setting and equal per-model hyperparameter search for all classical and deep baselines. CARDNet achieves the best held-out fleet accuracy with an RMSE of 1.040 Ah using only 53.7k parameters. Block-bootstrap confidence intervals separate CARDNet from all learned baselines on the held-out fleet, while leave-one-vehicle-out tests show statistically reliable gains over statistical forecasters and kernel regression and comparable mean accuracy to the strongest tree-ensemble, convolutional, and tuned attention baselines. Among this tied group, CARDNet provides superior stability and efficiency, achieving the lowest across-fold variance and worst-case error. On a single CPU core, it requires only 219 KiB of weights and 0.10 ms per inference, versus 3.5 MiB and 13.1 ms for the closest classical competitor. Error decomposition shows broad gains across week types, while a regeneration-weighted objective reduces regeneration-episode error by 5.2% at a 1.2% cost in overall accuracy.
- 11 November 2026
- Published by IOP Publishing in Engineering Research Express
CARDNet: A Lightweight Regeneration-Aware Network for ElectricVehicle Battery State-of-Health Estimation from Weekly Telematics Data
Accurate state-of-health (SOH) estimation is essential for safe and economical electric vehicle (EV) operation, yet real-world fleet telematics typically provides short, weekly aggregated sequences with pronounced capacity regeneration and substantial cross-vehicle heterogeneity, challenging deep models developed for dense laboratory data. This paper proposes CARDNet, a lightweight network that isolates high-frequency input fluctuations, where regeneration dynamics are concentrated, through a parameter-free moving-average decomposition. The extracted component is encoded by a compact convolutional branch and complemented by a global feature-interaction path that captures cross-feature nonlinearities. We further provide an explicit operational definition of capacity regeneration, under which the studied fleet exhibits 117 regeneration episodes that transiently reverse capacity equal to 48.8% of the net decay. Experiments on 20 real EVs monitored for over two years use a unified one-week-ahead prediction setting and equal per-model hyperparameter search for all classical and deep baselines. CARDNet achieves the best held-out fleet accuracy with an RMSE of 1.040 Ah using only 53.7k parameters. Block-bootstrap confidence intervals separate CARDNet from all learned baselines on the held-out fleet, while leave-one-vehicle-out tests show statistically reliable gains over statistical forecasters and kernel regression and comparable mean accuracy to the strongest tree-ensemble, convolutional, and tuned attention baselines. Among this tied group, CARDNet provides superior stability and efficiency, achieving the lowest across-fold variance and worst-case error. On a single CPU core, it requires only 219 KiB of weights and 0.10 ms per inference, versus 3.5 MiB and 13.1 ms for the closest classical competitor. Error decomposition shows broad gains across week types, while a regeneration-weighted objective reduces regeneration-episode error by 5.2% at a 1.2% cost in overall accuracy.
- 30 April 2026
- Published by IOP Publishing in Engineering Research Express
- Vol. 8 (9) , 095248
CSMS-YOLO: Object Detection Algorithm Based on Cross-Spatial Multi-Scale Feature Fusion for Foggy Weather
With the growing popularity of deep learning in autonomous driving and intelligent road traffic management, object detection under adverse weather has emerged as a key research focus in recent years. This study targets vehicles and pedestrians in foggy scenes and proposes a cross-spatial multi-scale feature fusion object detection algorithm, namely Cross-Spatial Multi-Scale feature fusion YOLO (CSMS-YOLO), based on the YOLOv8 framework. To tackle insufficient feature information of small objects in foggy environments, a Re-parameterized Convolution module (RepConv) is introduced to enhance feature extraction capability. Addressing weak small-object detection in existing algorithms, a Multi-Scale Head (MSHead) is proposed, integrating an adaptive weight fusion mechanism to mitigate multi-scale feature information loss and confusion. To strengthen feature representation, a Cross-Spatial Efficient Attention (CSEA) mechanism is designed to establish correlations among feature information across different spatial positions. For improved foggy-scene performance, the CUT method aids domain adaptation training, and a Super-Resolution Domain Classifier (SRDC) extends the algorithm’s applicability from clear to foggy weather. Experimental results show the model achieves mAP@0.5 of 38.2% on Foggy Cityscapes and 61.3% on RTTS datasets, effectively enhancing detection accuracy in foggy environments and providing a viable solution for unannotated foggy data. Finally, comparisons with mainstream object detection algorithms verify the effectiveness and advancement of the proposed CSMS-YOLO.
- 30 April 2026
- Published by IOP Publishing in Engineering Research Express
- Vol. 8 (9) , 095248
CSMS-YOLO: Object Detection Algorithm Based on Cross-Spatial Multi-Scale Feature Fusion for Foggy Weather
With the growing popularity of deep learning in autonomous driving and intelligent road traffic management, object detection under adverse weather has emerged as a key research focus in recent years. This study targets vehicles and pedestrians in foggy scenes and proposes a cross-spatial multi-scale feature fusion object detection algorithm, namely Cross-Spatial Multi-Scale feature fusion YOLO (CSMS-YOLO), based on the YOLOv8 framework. To tackle insufficient feature information of small objects in foggy environments, a Re-parameterized Convolution module (RepConv) is introduced to enhance feature extraction capability. Addressing weak small-object detection in existing algorithms, a Multi-Scale Head (MSHead) is proposed, integrating an adaptive weight fusion mechanism to mitigate multi-scale feature information loss and confusion. To strengthen feature representation, a Cross-Spatial Efficient Attention (CSEA) mechanism is designed to establish correlations among feature information across different spatial positions. For improved foggy-scene performance, the CUT method aids domain adaptation training, and a Super-Resolution Domain Classifier (SRDC) extends the algorithm’s applicability from clear to foggy weather. Experimental results show the model achieves mAP@0.5 of 38.2% on Foggy Cityscapes and 61.3% on RTTS datasets, effectively enhancing detection accuracy in foggy environments and providing a viable solution for unannotated foggy data. Finally, comparisons with mainstream object detection algorithms verify the effectiveness and advancement of the proposed CSMS-YOLO.
- Open Access
MulPViT-SimAM: An Electronic Substrate Defect Detection Framework for Addressing Class Imbalance Problems
As the cornerstone of contemporary electronics, the quality of electronic substrates—including Printed Circuit Boards (PCBs) and Ceramic Package Substrates (CPSs)—is intrinsic to product reliability. However, automated inspection is currently impeded by two persistent obstacles: the drastic multi-scale variation in defects and the acute class imbalance within defect datasets. Conventional deep learning approaches often fail to reconcile these challenges simultaneously, leading to suboptimal recognition of rare defect categories. To bridge this gap, we propose Multi-scale Partial Vision Transformer—Simple, Parameter-free Attention Module (MulPViT-SimAM), a robust framework designed for class-imbalanced electronic substrate defect detection. Our method features a novel multi-scale backbone (MulPViT) that synergizes partial convolutions with hierarchical attention mechanisms, facilitating the efficient extraction of both fine-grained local textures and global contextual dependencies. Additionally, we embed the Simple, Parameter-free Attention Module (SimAM) into the feature fusion stage to adaptively highlight defect-specific features while dampening background noise. To further mitigate data imbalance, we utilize the Equalized Focal Loss (EFL) function, which employs a category-specific modulating factor to dynamically equilibrate the learning focus across different classes. Comprehensive benchmarking reveals state-of-the-art performance, achieving mAP@0.5 scores of 95.7% on the standard PKU-MARKET-PCB dataset and 54.2% on the highly challenging CPS2D-AD dataset. Significantly, our approach effectively mitigates class imbalance, narrowing the performance deviation of rare categories to just 4.3% on the PKU-Market-PCB dataset and 1.4% on the CPS2D-AD dataset, compared to 11.8% and 7.5% in baseline models. These findings position MulPViT-SimAM as a viable and efficient solution for industrial quality control.
Open AccessMulPViT-SimAM: An Electronic Substrate Defect Detection Framework for Addressing Class Imbalance Problems
As the cornerstone of contemporary electronics, the quality of electronic substrates—including Printed Circuit Boards (PCBs) and Ceramic Package Substrates (CPSs)—is intrinsic to product reliability. However, automated inspection is currently impeded by two persistent obstacles: the drastic multi-scale variation in defects and the acute class imbalance within defect datasets. Conventional deep learning approaches often fail to reconcile these challenges simultaneously, leading to suboptimal recognition of rare defect categories. To bridge this gap, we propose Multi-scale Partial Vision Transformer—Simple, Parameter-free Attention Module (MulPViT-SimAM), a robust framework designed for class-imbalanced electronic substrate defect detection. Our method features a novel multi-scale backbone (MulPViT) that synergizes partial convolutions with hierarchical attention mechanisms, facilitating the efficient extraction of both fine-grained local textures and global contextual dependencies. Additionally, we embed the Simple, Parameter-free Attention Module (SimAM) into the feature fusion stage to adaptively highlight defect-specific features while dampening background noise. To further mitigate data imbalance, we utilize the Equalized Focal Loss (EFL) function, which employs a category-specific modulating factor to dynamically equilibrate the learning focus across different classes. Comprehensive benchmarking reveals state-of-the-art performance, achieving mAP@0.5 scores of 95.7% on the standard PKU-MARKET-PCB dataset and 54.2% on the highly challenging CPS2D-AD dataset. Significantly, our approach effectively mitigates class imbalance, narrowing the performance deviation of rare categories to just 4.3% on the PKU-Market-PCB dataset and 1.4% on the CPS2D-AD dataset, compared to 11.8% and 7.5% in baseline models. These findings position MulPViT-SimAM as a viable and efficient solution for industrial quality control.
- Open Access
- 12 March 2026
- Published by MDPI AG in Electronics
- Vol. 15 (6) , 1191
Toward Real-Time Industrial Small Object Inspection: Decoupled Attention and Multi-Scale Aggregation for PCB Defect Detection
PCB surface defect detection plays a critical role in ensuring electronics manufacturing quality. To address the challenges of small target defect detection, this study proposes PCB-YOLO, an enhanced lightweight detector based on YOLOv8n. PCB-YOLO introduces three key improvements. First, a RepViT-EMA Fusion Architecture (REFA) module is designed for deep backbone layers to strengthen feature extraction while suppressing background interference from complex circuit patterns. Second, a Multi-Scale Grouped Aggregation (MSGA) module is developed to reduce feature redundancy and improve spatial-semantic information extraction for multi-scale defects. Third, a Pixel-level Intersection over Union (PIoU) loss function is proposed to enable pixel-level IoU calculation with enhanced angular and area constraints for more precise localization. Extensive experiments on the PKU-Market-PCB dataset demonstrate that PCB-YOLO achieves 98.4% mAP@0.5, 97.4% recall, and 96.1% precision with only 2.4 M parameters, 6.9 G FLOPs, and an inference speed of 224 FPS, outperforming multiple state-of-the-art methods while maintaining real-time capability. Additional experiments on the DeepPCB dataset yield 99.0% mAP@0.5 and 80.4% mAP@0.5:0.95, confirming the cross-dataset generalization ability of the proposed method.
- 12 March 2026
- Published by MDPI AG in Electronics
- Vol. 15 (6) , 1191
Open AccessToward Real-Time Industrial Small Object Inspection: Decoupled Attention and Multi-Scale Aggregation for PCB Defect Detection
PCB surface defect detection plays a critical role in ensuring electronics manufacturing quality. To address the challenges of small target defect detection, this study proposes PCB-YOLO, an enhanced lightweight detector based on YOLOv8n. PCB-YOLO introduces three key improvements. First, a RepViT-EMA Fusion Architecture (REFA) module is designed for deep backbone layers to strengthen feature extraction while suppressing background interference from complex circuit patterns. Second, a Multi-Scale Grouped Aggregation (MSGA) module is developed to reduce feature redundancy and improve spatial-semantic information extraction for multi-scale defects. Third, a Pixel-level Intersection over Union (PIoU) loss function is proposed to enable pixel-level IoU calculation with enhanced angular and area constraints for more precise localization. Extensive experiments on the PKU-Market-PCB dataset demonstrate that PCB-YOLO achieves 98.4% mAP@0.5, 97.4% recall, and 96.1% precision with only 2.4 M parameters, 6.9 G FLOPs, and an inference speed of 224 FPS, outperforming multiple state-of-the-art methods while maintaining real-time capability. Additional experiments on the DeepPCB dataset yield 99.0% mAP@0.5 and 80.4% mAP@0.5:0.95, confirming the cross-dataset generalization ability of the proposed method.
- 8 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (5) , 8050-8063
SED-UAV: A Synergistic Framework of Lightweight Chaotic Encryption and Multiscale Feature Detection for Secure UAV Applications
The proliferation of uncrewed aerial vehicles (UAVs) in 6G-enabled edge computing presents a dual challenge of ensuring secure data transmission and performing accurate object detection on resource-constrained platforms. This article proposes a Synergistic Encryption and Detection framework (SED-UAV) that integrates a lightweight chaotic image encryption algorithm, lightweight, block-based, and improved chaotic image encryption algorithm (LB-ICA), and an enhanced small object detector, YOLO-LFP. The LB-ICA algorithm leverages an improved chaotic map and a plaintext-aware key mechanism using SHA-256. It achieves a key space of over $2^{100}$ and an average encryption speed of 20.3 ms per block, providing robust security against differential and chosen-plaintext attacks with high efficiency suitable for UAVs. For detection, the YOLO-LFP model enhances the YOLOv8 baseline with a hybrid attention module, an adaptive feature fusion strategy, and a dedicated small object head. Experimental results on the VisDrone2019 benchmark show YOLO-LFP achieves a state-of-the-art performance of 44.7% mAP@0.5, significantly outperforming existing models. This work provides a comprehensive solution for deploying secure, real-time computer vision applications on UAV platforms.
- 8 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (5) , 8050-8063
SED-UAV: A Synergistic Framework of Lightweight Chaotic Encryption and Multiscale Feature Detection for Secure UAV Applications
The proliferation of uncrewed aerial vehicles (UAVs) in 6G-enabled edge computing presents a dual challenge of ensuring secure data transmission and performing accurate object detection on resource-constrained platforms. This article proposes a Synergistic Encryption and Detection framework (SED-UAV) that integrates a lightweight chaotic image encryption algorithm, lightweight, block-based, and improved chaotic image encryption algorithm (LB-ICA), and an enhanced small object detector, YOLO-LFP. The LB-ICA algorithm leverages an improved chaotic map and a plaintext-aware key mechanism using SHA-256. It achieves a key space of over $2^{100}$ and an average encryption speed of 20.3 ms per block, providing robust security against differential and chosen-plaintext attacks with high efficiency suitable for UAVs. For detection, the YOLO-LFP model enhances the YOLOv8 baseline with a hybrid attention module, an adaptive feature fusion strategy, and a dedicated small object head. Experimental results on the VisDrone2019 benchmark show YOLO-LFP achieves a state-of-the-art performance of 44.7% mAP@0.5, significantly outperforming existing models. This work provides a comprehensive solution for deploying secure, real-time computer vision applications on UAV platforms.
- 24 February 2025
- Published by IOP Publishing in Engineering Research Express
- Vol. 7 (1) , 015263
HaMP: a proactive hash-based framework for manipulation localization
With the continuous advancement of image tampering techniques, traditional hash-based tampering detection algorithms face numerous challenges, particularly in handling complex tampering types and achieving object-level precise localization. Existing methods often rely on manually designed features, resulting in forensic hashes that lack sufficient detail and fail to capture the fine-grained information needed for accurate localization. To address these limitations, HaMP, a novel hash-based proactive tampering localization method, is proposed. HaMP introduces a Hash Construction Module (HCM) capable of generating compact yet information-rich image hashes. By integrating multi-scale feature extraction, the HCM produces robust global representations of the image. For precise tampering localization, the proposed Localization Module (LM) refines these representations by leveraging hash differences and multi-scale features. This enables accurate detection of tampered regions, even in the presence of subtle modifications. Experimental results demonstrate that HaMP significantly outperforms traditional methods, showcasing enhanced robustness, accuracy, and generalization across multiple datasets.
- 24 February 2025
- Published by IOP Publishing in Engineering Research Express
- Vol. 7 (1) , 015263
HaMP: a proactive hash-based framework for manipulation localization
With the continuous advancement of image tampering techniques, traditional hash-based tampering detection algorithms face numerous challenges, particularly in handling complex tampering types and achieving object-level precise localization. Existing methods often rely on manually designed features, resulting in forensic hashes that lack sufficient detail and fail to capture the fine-grained information needed for accurate localization. To address these limitations, HaMP, a novel hash-based proactive tampering localization method, is proposed. HaMP introduces a Hash Construction Module (HCM) capable of generating compact yet information-rich image hashes. By integrating multi-scale feature extraction, the HCM produces robust global representations of the image. For precise tampering localization, the proposed Localization Module (LM) refines these representations by leveraging hash differences and multi-scale features. This enables accurate detection of tampered regions, even in the presence of subtle modifications. Experimental results demonstrate that HaMP significantly outperforms traditional methods, showcasing enhanced robustness, accuracy, and generalization across multiple datasets.