Ying Wei
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
Ying Wei has a total of 137 co-authors. Here are the top co-authors by number of shared publications:
- Jiaguang Li
College of Information Science and Engineering, Northeastern University, Shenyang, China
9 shared publications Li-Ye Niu
College of Information Science and Engineering, Northeastern University, Shenyang, China
6 shared publications- Qian Hu
College of Information Science and Engineering, Northeastern University, Shenyang 110819, China
5 shared publications - Yuefeng Wang
College of Information Science and Engineering, Northeastern University, Shenyang, China
5 shared publications - Shuailei Ma
College of Information Science and Engineering, Northeastern University, Shenyang, China
4 shared publications - Xiang Li
College of Information Science and Engineering, Northeastern University, Shenyang 110819, China
4 shared publications
Latest Publications
- 30 September 2026
- Published by Springer Nature in Signal, Image and Video Processing
- Vol. 20 (13) , 673
EIQNet: extracting key support information from query images: a novel idea for few-shot segmentation
- 30 September 2026
- Published by Springer Nature in Signal, Image and Video Processing
- Vol. 20 (13) , 673
EIQNet: extracting key support information from query images: a novel idea for few-shot segmentation
- 16 September 2026
- Published by Elsevier in Neurocomputing
- Vol. 708
Matrix-based representation of mesh surface deformation for mesh-driven 3DGS
- 16 September 2026
- Published by Elsevier in Neurocomputing
- Vol. 708
Matrix-based representation of mesh surface deformation for mesh-driven 3DGS
- Open Access
- 9 June 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (12) , 2059
BAFNet: A Few-Shot Segmentation Algorithm Based on Two-Stage Backbone Network Adaptation Fine-Tuning via Meta-Learning
The objective of few-shot segmentation is to segment novel categories given only a few annotated support images. Current FSS methods typically rely on pretrained backbone networks while often overlooking the inherent discrepancy between pretraining tasks and downstream segmentation tasks. This oversight renders the models susceptible to noise interference and hinders rapid generalization to novel categories. To address these limitations, we propose BAFNet, a novel few-shot segmentation algorithm based on two-stage backbone adaptive fine-tuning. Our approach incorporates a Feature Activation Adapter module into the backbone network, which operates through similarity feature enhancement and low-dimensional adaptive learning. Building upon this foundation, we develop an adapter-based fine-tuning strategy for the training phase that enhances the backbone network’s capacity for extracting category-relevant features while optimizing similarity representation of the extracted features. Additionally, we introduce a support set-driven, in-episode, online fine-tuning strategy for the testing phase, which leverages data augmentation to generate pseudo-query sets for supervised fine-tuning optimization. Through comprehensive quantitative and qualitative experiments conducted on PASCAL-5i, COCO-20i, and the industrial MT Defect Dataset, our results demonstrate that the proposed BAFNet model achieves state-of-the-art few-shot segmentation performance while utilizing the minimal number of trainable parameters. Our method obtains superior performance for both the mean intersection over union and foreground-background intersection over union evaluation metrics, exhibiting remarkable applicability for both general images in complex scenes and industrial defect segmentation under few-shot conditions.
- 9 June 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (12) , 2059
Open AccessBAFNet: A Few-Shot Segmentation Algorithm Based on Two-Stage Backbone Network Adaptation Fine-Tuning via Meta-Learning
The objective of few-shot segmentation is to segment novel categories given only a few annotated support images. Current FSS methods typically rely on pretrained backbone networks while often overlooking the inherent discrepancy between pretraining tasks and downstream segmentation tasks. This oversight renders the models susceptible to noise interference and hinders rapid generalization to novel categories. To address these limitations, we propose BAFNet, a novel few-shot segmentation algorithm based on two-stage backbone adaptive fine-tuning. Our approach incorporates a Feature Activation Adapter module into the backbone network, which operates through similarity feature enhancement and low-dimensional adaptive learning. Building upon this foundation, we develop an adapter-based fine-tuning strategy for the training phase that enhances the backbone network’s capacity for extracting category-relevant features while optimizing similarity representation of the extracted features. Additionally, we introduce a support set-driven, in-episode, online fine-tuning strategy for the testing phase, which leverages data augmentation to generate pseudo-query sets for supervised fine-tuning optimization. Through comprehensive quantitative and qualitative experiments conducted on PASCAL-5i, COCO-20i, and the industrial MT Defect Dataset, our results demonstrate that the proposed BAFNet model achieves state-of-the-art few-shot segmentation performance while utilizing the minimal number of trainable parameters. Our method obtains superior performance for both the mean intersection over union and foreground-background intersection over union evaluation metrics, exhibiting remarkable applicability for both general images in complex scenes and industrial defect segmentation under few-shot conditions.
- Open Access
- 21 March 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (6) , 1067
JDC-DA: An Unsupervised Target Domain Algorithm for Alzheimer’s Disease Diagnosis with Structural MRI Using Joint Domain and Category Dual Adaptation
Domain shift in multi-source MRI imaging data significantly degrades the performance of Alzheimer’s disease diagnostic models. This study aims to develop an effective unsupervised domain adaptation method to enhance diagnostic accuracy across different clinical datasets. We propose a Joint Domain and Category Dual Adaptation framework (JDC-DA) that integrates metric learning and adversarial learning. The method employs multi-scale feature aggregation to capture diverse lesion characteristics, generates dynamic prototype features through category clustering, and implements a novel metric learning approach that simultaneously aligns both domain-level and category-level feature distributions. Additionally, we introduce a classification certainty maximization strategy that establishes a dual adversarial mechanism between domain discriminator and classification discrepancy discriminator. The framework was evaluated on four public datasets (ADNI-1, ADNI-2, ADNI-3, AIBL) containing 1230 baseline sMRI scans for four classification tasks: AD vs. NC, MCI vs. NC, AD vs. MCI, and AD vs. MCI vs. NC. The proposed JDC-DA method achieved superior performance with accuracies of 92.16%, 83.56%, 81.96%, and 79.12% for the four classification tasks respectively, significantly outperforming existing state-of-the-art domain adaptation methods across all evaluation metrics. The JDC-DA framework effectively addresses domain shift challenges in Alzheimer’s disease diagnosis through its integrated approach to feature alignment and adversarial learning. The method demonstrates strong potential for clinical application in automated diagnosis systems, particularly for handling multi-center neuroimaging data with distribution discrepancies.
- 21 March 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (6) , 1067
Open AccessJDC-DA: An Unsupervised Target Domain Algorithm for Alzheimer’s Disease Diagnosis with Structural MRI Using Joint Domain and Category Dual Adaptation
Domain shift in multi-source MRI imaging data significantly degrades the performance of Alzheimer’s disease diagnostic models. This study aims to develop an effective unsupervised domain adaptation method to enhance diagnostic accuracy across different clinical datasets. We propose a Joint Domain and Category Dual Adaptation framework (JDC-DA) that integrates metric learning and adversarial learning. The method employs multi-scale feature aggregation to capture diverse lesion characteristics, generates dynamic prototype features through category clustering, and implements a novel metric learning approach that simultaneously aligns both domain-level and category-level feature distributions. Additionally, we introduce a classification certainty maximization strategy that establishes a dual adversarial mechanism between domain discriminator and classification discrepancy discriminator. The framework was evaluated on four public datasets (ADNI-1, ADNI-2, ADNI-3, AIBL) containing 1230 baseline sMRI scans for four classification tasks: AD vs. NC, MCI vs. NC, AD vs. MCI, and AD vs. MCI vs. NC. The proposed JDC-DA method achieved superior performance with accuracies of 92.16%, 83.56%, 81.96%, and 79.12% for the four classification tasks respectively, significantly outperforming existing state-of-the-art domain adaptation methods across all evaluation metrics. The JDC-DA framework effectively addresses domain shift challenges in Alzheimer’s disease diagnosis through its integrated approach to feature alignment and adversarial learning. The method demonstrates strong potential for clinical application in automated diagnosis systems, particularly for handling multi-center neuroimaging data with distribution discrepancies.
- 3 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-15
Hierarchical Feature-Modulated Transformer With Introspective Interactions for Visual Grounding in Remote Sensing Images
Visual grounding (VG) in remote sensing (RS) images aims to localize objects referred to by a query text in complex aerial scenes. However, existing methods still suffer from limitations in multimodal interaction and alignment. Specifically, multiscale visual feature fusion often lacks effective textual semantic filtering, introducing redundant background information that degrades cross-modal alignment. Moreover, although bidirectional cross-attention enhances intermodal information exchange, it tends to neglect intramodal semantic consistency, resulting in unstable and less discriminative representations. In addition, language queries obtained after cross-modal interaction are vulnerable to linguistic noise during decoding, which further impairs localization robustness. To address these challenges in a unified and progressive manner, we propose a hierarchical vision–language alignment framework, termed hierarchical feature-modulated transformer with introspective interaction (HMIT). Specifically, text-modulated cross-scale feature fusion (TMCFF) performs early-stage text-guided modulation on multiscale visual features to suppress text-irrelevant background information. Cross-modal interaction with contextual introspection (CMICI) refines cross-modal representations through contextual introspection while preserving intramodal semantic consistency. vision-guided text calibration (VGTC) calibrates high-level language queries using global visual priors prior to decoding, producing more robust and discriminative query embeddings. Extensive experiments on three benchmark datasets demonstrate that HMIT consistently outperforms existing methods across multiple metrics, achieving new state-of-the-art performance.
- 3 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-15
Hierarchical Feature-Modulated Transformer With Introspective Interactions for Visual Grounding in Remote Sensing Images
Visual grounding (VG) in remote sensing (RS) images aims to localize objects referred to by a query text in complex aerial scenes. However, existing methods still suffer from limitations in multimodal interaction and alignment. Specifically, multiscale visual feature fusion often lacks effective textual semantic filtering, introducing redundant background information that degrades cross-modal alignment. Moreover, although bidirectional cross-attention enhances intermodal information exchange, it tends to neglect intramodal semantic consistency, resulting in unstable and less discriminative representations. In addition, language queries obtained after cross-modal interaction are vulnerable to linguistic noise during decoding, which further impairs localization robustness. To address these challenges in a unified and progressive manner, we propose a hierarchical vision–language alignment framework, termed hierarchical feature-modulated transformer with introspective interaction (HMIT). Specifically, text-modulated cross-scale feature fusion (TMCFF) performs early-stage text-guided modulation on multiscale visual features to suppress text-irrelevant background information. Cross-modal interaction with contextual introspection (CMICI) refines cross-modal representations through contextual introspection while preserving intramodal semantic consistency. vision-guided text calibration (VGTC) calibrates high-level language queries using global visual priors prior to decoding, producing more robust and discriminative query embeddings. Extensive experiments on three benchmark datasets demonstrate that HMIT consistently outperforms existing methods across multiple metrics, achieving new state-of-the-art performance.
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 33 (10709908) , 376-380
Explicit-Implicit Prompt Injection and Semantic-Guided Latent LoRA for Vision-Language Tracking
Prompt-based learning has shown promise in visual-language tracking (VLT), yet existing methods often rely on either explicit or implicit prompting alone, limiting fine-grained cross-modal alignment. Moreover, Low-Rank Adaptation (LoRA) -based fine-tuning in prior work typically focuses on visual-only adaptation, overlooking language semantics. To address these issues, we propose a unified VLT framework that integrates Explicit-Implicit Prompt Injection (EIPI) and Semantic-Guided Latent LoRA (SGLL). EIPI introduces semantic prompts to facilitate robust and context-sensitive target modeling through two pathways. The explicit prompts are constructed by interact between multi-modal target representations with the search region, while implicit prompts are learned from linguistic features via a lightweight bottleneck network. Then, SGLL extends standard LoRA by introducing learnable queries in the latent space, allowing residual modulation based on language-visual semantics without retraining the full model. This dual design yields a parameter-efficient tracker with strong cross-modal adaptability. Extensive experiments show our method outperforms prior prompt-based approaches while maintaining high efficiency.
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 33 (10709908) , 376-380
Explicit-Implicit Prompt Injection and Semantic-Guided Latent LoRA for Vision-Language Tracking
Prompt-based learning has shown promise in visual-language tracking (VLT), yet existing methods often rely on either explicit or implicit prompting alone, limiting fine-grained cross-modal alignment. Moreover, Low-Rank Adaptation (LoRA) -based fine-tuning in prior work typically focuses on visual-only adaptation, overlooking language semantics. To address these issues, we propose a unified VLT framework that integrates Explicit-Implicit Prompt Injection (EIPI) and Semantic-Guided Latent LoRA (SGLL). EIPI introduces semantic prompts to facilitate robust and context-sensitive target modeling through two pathways. The explicit prompts are constructed by interact between multi-modal target representations with the search region, while implicit prompts are learned from linguistic features via a lightweight bottleneck network. Then, SGLL extends standard LoRA by introducing learnable queries in the latent space, allowing residual modulation based on language-visual semantics without retraining the full model. This dual design yields a parameter-efficient tracker with strong cross-modal adaptability. Extensive experiments show our method outperforms prior prompt-based approaches while maintaining high efficiency.
- 9 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 7 (3) , 1803-1815
FCA-HLP: Multilayer Feature Cross-Activation Network With High- and Low-Level Prototypes for Few-Shot Segmentation
Few-shot segmentation (FSS) strives to segment novel categories using merely a limited number of labeled images. Current prototype learning and correlation learning approaches struggle to effectively harness both high- and low-level information from support and query images, leading to suboptimal segmentation results. In this work, we propose a multi-layer feature cross-activation (FCA) network with high- and low-level prototypes (HLP), which fully utilizes support and query information from both features and prototypes perspectives. Specifically, for the FCA module, we design a simple activation method that uses all the pixel-level support foreground features to activate query features, thereby obtaining activation maps without losing pixel-level detail information of support features. For the HLP module, we combine image-level prototypes with pixel-level prototypes to fully utilize high-level category information and low-level attribute information of support features. Besides, the prototype generation method integrates query information, which further enables the prototypes to better match target query features, especially when there are great differences between query and support images. Extensive experiments on PASCAL-5i and COCO-20i under 1-shot and 5-shot settings validate the effectiveness of our FCA-HLP. Our method establishes new state-of-the-art performance. Additionally, we analyze the performance of the multi-layer feature cross-activation network in the absence of the HLP module. The results indicate that even without prototypes, the FCA module can still deliver strong performance. Codes are available at https://github.com/Jiaguang-NEU/FCA-HLP.
- 9 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 7 (3) , 1803-1815
FCA-HLP: Multilayer Feature Cross-Activation Network With High- and Low-Level Prototypes for Few-Shot Segmentation
Few-shot segmentation (FSS) strives to segment novel categories using merely a limited number of labeled images. Current prototype learning and correlation learning approaches struggle to effectively harness both high- and low-level information from support and query images, leading to suboptimal segmentation results. In this work, we propose a multi-layer feature cross-activation (FCA) network with high- and low-level prototypes (HLP), which fully utilizes support and query information from both features and prototypes perspectives. Specifically, for the FCA module, we design a simple activation method that uses all the pixel-level support foreground features to activate query features, thereby obtaining activation maps without losing pixel-level detail information of support features. For the HLP module, we combine image-level prototypes with pixel-level prototypes to fully utilize high-level category information and low-level attribute information of support features. Besides, the prototype generation method integrates query information, which further enables the prototypes to better match target query features, especially when there are great differences between query and support images. Extensive experiments on PASCAL-5i and COCO-20i under 1-shot and 5-shot settings validate the effectiveness of our FCA-HLP. Our method establishes new state-of-the-art performance. Additionally, we analyze the performance of the multi-layer feature cross-activation network in the absence of the HLP module. The results indicate that even without prototypes, the FCA module can still deliver strong performance. Codes are available at https://github.com/Jiaguang-NEU/FCA-HLP.
- 19 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (12) , 11738-11752
SKDF: A Simple Knowledge Distillation Framework for Distilling Open-Vocabulary Knowledge to Open-World Object Detector
Open World Object Detection (OWOD) is a novel computer vision task with a considerable challenge, bridging the gap between classic object detection (OD) and real-world object detection. In addition to detecting and classifying seen/known objects, OWOD algorithms are expected to localize all potential unseen/unknown objects and incrementally learn them. The large pre-trained vision-language grounding models (VLM, e.g., GLIP) have rich knowledge about the open world, but are limited by text prompts and cannot localize indescribable objects. However, there are many detection scenarios in which pre-defined language descriptions are unavailable during inference. In this paper, we attempt to specialize the VLM model for OWOD tasks by distilling its open-world knowledge into a language-agnostic detector. Surprisingly, we observe that the simple knowledge distillation approach leads to unexpected performance for unknown object detection, even with a small amount of data. Unfortunately, knowledge distillation for unknown objects severely affects the learning of detectors with conventional structures, leading to catastrophic damage to the model’s ability to learn about known objects. To alleviate these problems, we propose the down-weight training strategy for knowledge distillation from vision-language model to single visual modality one. Meanwhile, we propose the cascade decoupled decoders that decouple the learning of localization and recognition to reduce the impact of category interactions of known and unknown objects on the localization learning process. Ablation experiments demonstrate that both of them are effective in mitigating the impact of open-world knowledge distillation on the learning of known objects. Additionally, to alleviate the current lack of comprehensive benchmarks for evaluating the ability of the open-world detector to detect unknown objects in the open world, we refine the benchmark for evaluating the performance of unknown object detection by augmenting annotations for unknown objects which we name“IntensiveSet$\scriptstyle\spadesuit$ ”. Comprehensive experiments performed on OWOD, MS-COCO, and our proposed benchmarks demonstrate the effectiveness of our methods.- 19 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (12) , 11738-11752
SKDF: A Simple Knowledge Distillation Framework for Distilling Open-Vocabulary Knowledge to Open-World Object Detector
Open World Object Detection (OWOD) is a novel computer vision task with a considerable challenge, bridging the gap between classic object detection (OD) and real-world object detection. In addition to detecting and classifying seen/known objects, OWOD algorithms are expected to localize all potential unseen/unknown objects and incrementally learn them. The large pre-trained vision-language grounding models (VLM, e.g., GLIP) have rich knowledge about the open world, but are limited by text prompts and cannot localize indescribable objects. However, there are many detection scenarios in which pre-defined language descriptions are unavailable during inference. In this paper, we attempt to specialize the VLM model for OWOD tasks by distilling its open-world knowledge into a language-agnostic detector. Surprisingly, we observe that the simple knowledge distillation approach leads to unexpected performance for unknown object detection, even with a small amount of data. Unfortunately, knowledge distillation for unknown objects severely affects the learning of detectors with conventional structures, leading to catastrophic damage to the model’s ability to learn about known objects. To alleviate these problems, we propose the down-weight training strategy for knowledge distillation from vision-language model to single visual modality one. Meanwhile, we propose the cascade decoupled decoders that decouple the learning of localization and recognition to reduce the impact of category interactions of known and unknown objects on the localization learning process. Ablation experiments demonstrate that both of them are effective in mitigating the impact of open-world knowledge distillation on the learning of known objects. Additionally, to alleviate the current lack of comprehensive benchmarks for evaluating the ability of the open-world detector to detect unknown objects in the open world, we refine the benchmark for evaluating the performance of unknown object detection by augmenting annotations for unknown objects which we name“IntensiveSet$\scriptstyle\spadesuit$ ”. Comprehensive experiments performed on OWOD, MS-COCO, and our proposed benchmarks demonstrate the effectiveness of our methods.- 31 March 2025
- Published by Springer Nature in Earth Science Informatics
- Vol. 18 (2) , 1-18
Segmentation of urban green space and water body based on high-resolution remote sensing images
- 31 March 2025
- Published by Springer Nature in Earth Science Informatics
- Vol. 18 (2) , 1-18
Segmentation of urban green space and water body based on high-resolution remote sensing images
- 13 February 2025
- Published by Elsevier in Image and Vision Computing
- Vol. 155
ESDA: Zero-shot semantic segmentation based on an embedding semantic space distribution adjustment strategy
- 13 February 2025
- Published by Elsevier in Image and Vision Computing
- Vol. 155
ESDA: Zero-shot semantic segmentation based on an embedding semantic space distribution adjustment strategy