Guangwei Zhang
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
Guangwei Zhang has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Bing Zhou
The College of Computer Science and Artificial Intelligence, Zhengzhou University, Zhengzhou, China
3 shared publications - Xiangqian Liu
The College of Computer Science and Artificial Intelligence, Zhengzhou University, Zhengzhou, China
3 shared publications - Yifan Zhu
Beijing University of Posts and Telecommunications, Beijing, China
3 shared publications - Fangxiang Feng
Beijing University of Posts and Telecommunications, Beijing, China
2 shared publications - Haoran Luo
Nanyang technological university, Singapore, Singapore
2 shared publications - Meina Song
Beijing University of Posts and Telecommunications, Beijing, China
2 shared publications
Latest Publications
- 2 August 2026
- Published by Springer Nature
- p. 3-15
Cross-Layer Feature Aggregation Object Detection Networks for Modality-enhanced UAV Images
- 2 August 2026
- Published by Springer Nature
- p. 3-15
Cross-Layer Feature Aggregation Object Detection Networks for Modality-enhanced UAV Images
- 25 July 2026
- Published by Springer Nature in The Visual Computer
- Vol. 42 (10) , 434
TSMMR: a mixed Mamba network with cross-drone template fusion and text-guided redetection for robust aerial tracking
- 25 July 2026
- Published by Springer Nature in The Visual Computer
- Vol. 42 (10) , 434
TSMMR: a mixed Mamba network with cross-drone template fusion and text-guided redetection for robust aerial tracking
- 30 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (8) , 16434-16450
Dynamic Spatiotemporal Information Interaction for Multidrone Single Object Tracking
Multidrone single object tracking is a key technology in the Internet of Things (IoT)-based aerial sensing systems. However, existing methods often neglect both the temporal continuity of the object across frames and spatial complementarity from multidrone viewpoints, limiting their robustness in large-scale and dynamic IoT environments. To address these issues, this article proposes a DSTII-MDOT, where spatiotemporal information refers to the object’s feature evolution over time (temporal) and its complementary multiview representations (spatial). First, a dynamic temporal feature aggregation (DTFA) module is proposed, which captures the object’s motion and appearance variations by exploiting frame-to-frame differences, thereby ensuring feature continuity and stable predictions across frames in long-term tracking. Next, a spatial alignment method for cross-modal (text–image) semantic alignment (CMSA) is proposed. This method ensures the semantic alignment between visual features and textual descriptions by leveraging the contrastive language-image pretraining (CLIP) model and employs a multihead cross-attention mechanism to capture the Top- $K$ regions of interest within the search area. This enhances cross-drone collaboration and reduces redundant communications in the IoT networks. Finally, a wavelet frequency-domain feature refinement (WFDFR) module is proposed to enhance the texture features of the template image, effectively solving the problem of blurred texture features or missing details in complex scenes. The experimental results on the MDOT dataset demonstrate that the proposed DSTII-MDOT tracker surpasses existing advanced methods in both success rate (SR) and precision, validating the effectiveness and superiority of the proposed method. The code and models are available at https://github.com/JerryBryant24/DSTII-MDOT.
- 30 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (8) , 16434-16450
Dynamic Spatiotemporal Information Interaction for Multidrone Single Object Tracking
Multidrone single object tracking is a key technology in the Internet of Things (IoT)-based aerial sensing systems. However, existing methods often neglect both the temporal continuity of the object across frames and spatial complementarity from multidrone viewpoints, limiting their robustness in large-scale and dynamic IoT environments. To address these issues, this article proposes a DSTII-MDOT, where spatiotemporal information refers to the object’s feature evolution over time (temporal) and its complementary multiview representations (spatial). First, a dynamic temporal feature aggregation (DTFA) module is proposed, which captures the object’s motion and appearance variations by exploiting frame-to-frame differences, thereby ensuring feature continuity and stable predictions across frames in long-term tracking. Next, a spatial alignment method for cross-modal (text–image) semantic alignment (CMSA) is proposed. This method ensures the semantic alignment between visual features and textual descriptions by leveraging the contrastive language-image pretraining (CLIP) model and employs a multihead cross-attention mechanism to capture the Top- $K$ regions of interest within the search area. This enhances cross-drone collaboration and reduces redundant communications in the IoT networks. Finally, a wavelet frequency-domain feature refinement (WFDFR) module is proposed to enhance the texture features of the template image, effectively solving the problem of blurred texture features or missing details in complex scenes. The experimental results on the MDOT dataset demonstrate that the proposed DSTII-MDOT tracker surpasses existing advanced methods in both success rate (SR) and precision, validating the effectiveness and superiority of the proposed method. The code and models are available at https://github.com/JerryBryant24/DSTII-MDOT.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71 (2) , 4714-4720
Intelligent Decision Support Systems for Energy-Efficient Autonomous Systems in Consumer Electronics
This paper proposes an AI-based decision support system for enhancing energy efficiency in autonomous systems within consumer electronics. The primary objective is to optimize energy consumption by leveraging advanced Recurrent Neural Networks (RNNs) combined with Adaptive Moment Estimation (Adam). The proposed intelligent framework incorporates real-time data processing, prediction, and adaptive decision-making to reduce energy usage across a variety of devices, such as smart thermostats, robotic vacuum cleaners, and IoT-enabled lighting systems. Key methodologies include the integration of RNNs to model temporal energy consumption patterns and the Adam optimizer for efficient model training. The results demonstrate significant energy savings with improvements in prediction accuracy compared to traditional approaches. The findings suggest that the AI-based system can provide substantial reductions in energy consumption while maintaining or improving device performance. The research has important implications for the future of energy-efficient consumer electronics, particularly as the demand for smart devices grows.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 71 (2) , 4714-4720
Intelligent Decision Support Systems for Energy-Efficient Autonomous Systems in Consumer Electronics
This paper proposes an AI-based decision support system for enhancing energy efficiency in autonomous systems within consumer electronics. The primary objective is to optimize energy consumption by leveraging advanced Recurrent Neural Networks (RNNs) combined with Adaptive Moment Estimation (Adam). The proposed intelligent framework incorporates real-time data processing, prediction, and adaptive decision-making to reduce energy usage across a variety of devices, such as smart thermostats, robotic vacuum cleaners, and IoT-enabled lighting systems. Key methodologies include the integration of RNNs to model temporal energy consumption patterns and the Adam optimizer for efficient model training. The results demonstrate significant energy savings with improvements in prediction accuracy compared to traditional approaches. The findings suggest that the AI-based system can provide substantial reductions in energy consumption while maintaining or improving device performance. The research has important implications for the future of energy-efficient consumer electronics, particularly as the demand for smart devices grows.
- Open Access
- 22 April 2025
- Published by Association for Computing Machinery (ACM)
- p. 134-146
PM-MOE: Mixture of Experts on Private Model Parameters for Personalized Federated Learning
- 22 April 2025
- Published by Association for Computing Machinery (ACM)
- p. 134-146
Open AccessPM-MOE: Mixture of Experts on Private Model Parameters for Personalized Federated Learning
- 26 February 2025
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 81 (4) , 1-26
An efficient feature aggregation network for small object detection in UAV aerial images
- 26 February 2025
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 81 (4) , 1-26
An efficient feature aggregation network for small object detection in UAV aerial images
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 2729-2738
CP-Prompt: Composition-Based Cross-modal Prompting for Domain-Incremental Continual Learning
The key challenge of cross-modal domain-incremental learning (DIL) is to enable the learning model to continuously learn from novel data with different feature distributions under the same task without forgetting old ones. However, existing top-performing methods still cause high forgetting rates, by lacking intra-domain knowledge extraction and inter-domain common prompting strategy. In this paper, we propose a simple yet effective framework, CP-Prompt, by training limited parameters to instruct a pre-trained model to learn new domains and avoid forgetting existing feature distributions. CP-Prompt captures intra-domain knowledge by compositionally inserting personalized prompts into multi-head self-attention layers and then learns the inter-domain knowledge with a common prompting strategy. CP-Prompt shows superiority compared with state-of-the-art baselines among three widely evaluated DIL tasks. The source code is available at https://github.com/dannis97500/CP_Prompt.
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 2729-2738
Open AccessCP-Prompt: Composition-Based Cross-modal Prompting for Domain-Incremental Continual Learning
The key challenge of cross-modal domain-incremental learning (DIL) is to enable the learning model to continuously learn from novel data with different feature distributions under the same task without forgetting old ones. However, existing top-performing methods still cause high forgetting rates, by lacking intra-domain knowledge extraction and inter-domain common prompting strategy. In this paper, we propose a simple yet effective framework, CP-Prompt, by training limited parameters to instruct a pre-trained model to learn new domains and avoid forgetting existing feature distributions. CP-Prompt captures intra-domain knowledge by compositionally inserting personalized prompts into multi-head self-attention layers and then learns the inter-domain knowledge with a common prompting strategy. CP-Prompt shows superiority compared with state-of-the-art baselines among three widely evaluated DIL tasks. The source code is available at https://github.com/dannis97500/CP_Prompt.
- Open Access
- 20 August 2023
- Published by MDPI AG in Electronics
- Vol. 12 (16) , 3521
A Visually Enhanced Neural Encoder for Synset Induction
The synset induction task is to automatically cluster semantically identical instances, which are often represented by texts and images. Previous works mainly consider textual parts, while ignoring the visual counterparts. However, how to effectively employ the visual information to enhance the semantic representation for the synset induction is challenging. In this paper, we propose a Visually Enhanced NeUral Encoder (i.e., VENUE) to learn a multimodal representation for the synset induction task. The key insight lies in how to construct multimodal representations through intra-modal and inter-modal interactions among images and text. Specifically, we first design the visual interaction module through the attention mechanism to capture the correlation among images. To obtain the multi-granularity textual representations, we fuse the pre-trained tags and word embeddings. Second, we design a masking module to filter out weakly relevant visual information. Third, we present a gating module to adaptively regulate the modalities’ contributions to semantics. A triplet loss is adopted to train the VENUE encoder for learning discriminative multimodal representations. Then, we perform clustering algorithms on the obtained representations to induce synsets. To verify our approach, we collect a multimodal dataset, i.e., MMAI-Synset, and conduct extensive experiments. The experimental results demonstrate that our method outperforms strong baselines on three groups of evaluation metrics.
- 20 August 2023
- Published by MDPI AG in Electronics
- Vol. 12 (16) , 3521
Open AccessA Visually Enhanced Neural Encoder for Synset Induction
The synset induction task is to automatically cluster semantically identical instances, which are often represented by texts and images. Previous works mainly consider textual parts, while ignoring the visual counterparts. However, how to effectively employ the visual information to enhance the semantic representation for the synset induction is challenging. In this paper, we propose a Visually Enhanced NeUral Encoder (i.e., VENUE) to learn a multimodal representation for the synset induction task. The key insight lies in how to construct multimodal representations through intra-modal and inter-modal interactions among images and text. Specifically, we first design the visual interaction module through the attention mechanism to capture the correlation among images. To obtain the multi-granularity textual representations, we fuse the pre-trained tags and word embeddings. Second, we design a masking module to filter out weakly relevant visual information. Third, we present a gating module to adaptively regulate the modalities’ contributions to semantics. A triplet loss is adopted to train the VENUE encoder for learning discriminative multimodal representations. Then, we perform clustering algorithms on the obtained representations to induce synsets. To verify our approach, we collect a multimodal dataset, i.e., MMAI-Synset, and conduct extensive experiments. The experimental results demonstrate that our method outperforms strong baselines on three groups of evaluation metrics.
- Open Access
HeadText: Exploring Hands-free Text Entry using Head Gestures by Motion Sensing on a Smart Earpiece
We present HeadText, a hands-free technique on a smart earpiece for text entry by motion sensing. Users input text utilizing only 7 head gestures for key selection, word selection, word commitment and word cancelling tasks. Head gesture recognition is supported by motion sensing on a smart earpiece to capture head moving signals and machine learning algorithms (K-Nearest-Neighbor (KNN) with a Dynamic Time Warping (DTW) distance measurement). A 10-participant user study proved that HeadText could recognize 7 head gestures at an accuracy of 94.29%. After that, the second user study presented that HeadText could achieve a maximum accuracy of 10.65 WPM and an average accuracy of 9.84 WPM for text entry. Finally, we demonstrate potential applications of HeadText in hands-free scenarios for (a). text entry of people with motor impairments, (b). private text entry, and (c). socially acceptable text entry.
Open AccessHeadText: Exploring Hands-free Text Entry using Head Gestures by Motion Sensing on a Smart Earpiece
We present HeadText, a hands-free technique on a smart earpiece for text entry by motion sensing. Users input text utilizing only 7 head gestures for key selection, word selection, word commitment and word cancelling tasks. Head gesture recognition is supported by motion sensing on a smart earpiece to capture head moving signals and machine learning algorithms (K-Nearest-Neighbor (KNN) with a Dynamic Time Warping (DTW) distance measurement). A 10-participant user study proved that HeadText could recognize 7 head gestures at an accuracy of 94.29%. After that, the second user study presented that HeadText could achieve a maximum accuracy of 10.65 WPM and an average accuracy of 9.84 WPM for text entry. Finally, we demonstrate potential applications of HeadText in hands-free scenarios for (a). text entry of people with motor impairments, (b). private text entry, and (c). socially acceptable text entry.
- 10 February 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 22 (12) , 3075-3087
Exploring Global and Local Linguistic Representations for Text-to-Image Synthesis
The task of text-to-image synthesis is to generate photographic images conditioned on given textual descriptions. This challenging task has recently attracted considerable attention from the multimedia community due to its potential applications. Most of the up-to-date approaches are built based on generative adversarial network (GAN) models, and they synthesize images conditioned on the global linguistic representation. However, the sparsity of the global representation results in training difficulties on GANs and a shortage of fine-grained information in the generated images. To address this problem, we propose cross-modal global and local linguistic representations-based generative adversarial networks (CGL-GAN) by incorporating the local linguistic representation into the GAN. In our CGL-GAN, we construct a generator to synthesize the target images and a discriminator to judge whether the generated images conform with the text description. In the discriminator, we construct the cross-modal correlation by projecting the image representations at high and low levels onto the global and local linguistic representations, respectively. We design the hinge loss function to train our CGL-GAN model. We evaluate the proposed CGL-GAN on two publicly available datasets, the CUB and the MS-COCO. The extensive experiments demonstrate that incorporating fine-grained local linguistic information with cross-modal correlation can greatly improve the performance of text-to-image synthesis, even when generating high-resolution images.
- 10 February 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 22 (12) , 3075-3087
Exploring Global and Local Linguistic Representations for Text-to-Image Synthesis
The task of text-to-image synthesis is to generate photographic images conditioned on given textual descriptions. This challenging task has recently attracted considerable attention from the multimedia community due to its potential applications. Most of the up-to-date approaches are built based on generative adversarial network (GAN) models, and they synthesize images conditioned on the global linguistic representation. However, the sparsity of the global representation results in training difficulties on GANs and a shortage of fine-grained information in the generated images. To address this problem, we propose cross-modal global and local linguistic representations-based generative adversarial networks (CGL-GAN) by incorporating the local linguistic representation into the GAN. In our CGL-GAN, we construct a generator to synthesize the target images and a discriminator to judge whether the generated images conform with the text description. In the discriminator, we construct the cross-modal correlation by projecting the image representations at high and low levels onto the global and local linguistic representations, respectively. We design the hinge loss function to train our CGL-GAN model. We evaluate the proposed CGL-GAN on two publicly available datasets, the CUB and the MS-COCO. The extensive experiments demonstrate that incorporating fine-grained local linguistic information with cross-modal correlation can greatly improve the performance of text-to-image synthesis, even when generating high-resolution images.