Yang 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
Yang Zhang has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
Hongbo Yang
School of Automation, Beijing Information Science and Technology University, Beijing, People’s Republic of China
3 shared publicationsQingjuan Feng
School of Automation, Beijing Information Science and Technology University, Beijing, China
3 shared publicationsKebei Zhang
Beijing Institute of Control Engineering, Beijing, China
2 shared publications- Wei Liu
School of Automation, Beijing Information Science and Technology University, Beijing, China
2 shared publications Zexi Deng
Beijing Information Science and Technology University, School of Automation, Beijing, China
2 shared publications- Bin Gu
China Acad Elect & Informat Technol, Beijing, Peoples R China
1 shared publication
Latest Publications
- 10 March 2026
- Published by Springer Nature in Multimedia Systems
- Vol. 32 (3) , 191
CSPGNet: few-shot segmentation in X-ray image based on cross-attention and self-adaptive prototype guidance
- 10 March 2026
- Published by Springer Nature in Multimedia Systems
- Vol. 32 (3) , 191
CSPGNet: few-shot segmentation in X-ray image based on cross-attention and self-adaptive prototype guidance
- Open Access
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 18 (19391404) , 25845-25862
SD-DETR: Space Debris Detection Transformer Based on Dynamic Convolutional Network and Cross-Scale Collaborative Attention
With the advancement of human exploration of near-Earth space, the number of space debris in Earth's orbit has increased dramatically, posing serious threats to orbiting spacecraft. Traditional space debris detection methods exhibit significant limitations in handling small-scale targets, complex backgrounds, and real-time requirements. This paper proposes the Space Debris Detection Transformer (SD-DETR) algorithm, an end-to-end detector specifically optimized for space debris detection tasks. First, we design a Dynamic Convolutional Network backbone that enables adaptive feature extraction for multi-scale, irregularly shaped debris through the dynamic convolution mixing mechanism of the Adaptive Multi-scale Feature Block. Second, we propose a Cross-scale Collaborative Attention Fusion Module that realizes deep feature fusion through hybrid domain feature enhancement and channel-guided spatial attention mechanisms. Finally, we introduce a Multi-Level Adaptive Feature Bridging module that enhances cross-hierarchical feature integration through dual-scale spatial perception attention. On the constructed Nebula space debris dataset, SD-DETR achieves 89.4% mAP50 and 75.4% mAP50-95, improving by 2.1% and 2.9% respectively compared to the baseline model. Generalization experiments on the NCSTP-10000 dataset further validate the effectiveness of the method, demonstrating its superior detection performance in complex space environments.
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 18 (19391404) , 25845-25862
Open AccessSD-DETR: Space Debris Detection Transformer Based on Dynamic Convolutional Network and Cross-Scale Collaborative Attention
With the advancement of human exploration of near-Earth space, the number of space debris in Earth's orbit has increased dramatically, posing serious threats to orbiting spacecraft. Traditional space debris detection methods exhibit significant limitations in handling small-scale targets, complex backgrounds, and real-time requirements. This paper proposes the Space Debris Detection Transformer (SD-DETR) algorithm, an end-to-end detector specifically optimized for space debris detection tasks. First, we design a Dynamic Convolutional Network backbone that enables adaptive feature extraction for multi-scale, irregularly shaped debris through the dynamic convolution mixing mechanism of the Adaptive Multi-scale Feature Block. Second, we propose a Cross-scale Collaborative Attention Fusion Module that realizes deep feature fusion through hybrid domain feature enhancement and channel-guided spatial attention mechanisms. Finally, we introduce a Multi-Level Adaptive Feature Bridging module that enhances cross-hierarchical feature integration through dual-scale spatial perception attention. On the constructed Nebula space debris dataset, SD-DETR achieves 89.4% mAP50 and 75.4% mAP50-95, improving by 2.1% and 2.9% respectively compared to the baseline model. Generalization experiments on the NCSTP-10000 dataset further validate the effectiveness of the method, demonstrating its superior detection performance in complex space environments.
- 1 November 2024
- Published by SPIE-Intl Soc Optical Eng in Journal of Electronic Imaging
- Vol. 33 (06) , 063055-17
Efficient video transmission using object segmentation over orthogonal frequency-division multiplexing-based underwater acoustic channel
The increase in transmission errors poses a serious challenge to underwater acoustic (UWA) communication methods, especially with the real-time error-correcting transmission requirements for underwater videos. We propose a video transmission method that enhances efficiency by using target segmentation. By integrating the reliability of UWA channels with the structure of compressed video, this approach effectively combats the loss of crucial information. An enhanced YOLOv5 algorithm is proposed to partition video information flow into different importance levels, and orthogonal frequency-division multiplexing (OFDM) is employed for transmission. This allows video data assigned to higher importance levels to be transmitted over the low-frequency bands with lower attenuation, whereas less important data are transmitted over the high-frequency bands with higher attenuation. Simulation results demonstrate that our proposed underwater video transmission method significantly improves the peak signal-to-noise ratio compared with three baseline techniques in the signal-to-noise ratio range of 25 to 50.
- 1 November 2024
- Published by SPIE-Intl Soc Optical Eng in Journal of Electronic Imaging
- Vol. 33 (06) , 063055-17
Efficient video transmission using object segmentation over orthogonal frequency-division multiplexing-based underwater acoustic channel
The increase in transmission errors poses a serious challenge to underwater acoustic (UWA) communication methods, especially with the real-time error-correcting transmission requirements for underwater videos. We propose a video transmission method that enhances efficiency by using target segmentation. By integrating the reliability of UWA channels with the structure of compressed video, this approach effectively combats the loss of crucial information. An enhanced YOLOv5 algorithm is proposed to partition video information flow into different importance levels, and orthogonal frequency-division multiplexing (OFDM) is employed for transmission. This allows video data assigned to higher importance levels to be transmitted over the low-frequency bands with lower attenuation, whereas less important data are transmitted over the high-frequency bands with higher attenuation. Simulation results demonstrate that our proposed underwater video transmission method significantly improves the peak signal-to-noise ratio compared with three baseline techniques in the signal-to-noise ratio range of 25 to 50.
- 24 October 2024
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (1) , 015111
Underwater point cloud transmission framework: hybrid encoder implementation based on CNN and transformer
In underwater multi-robot systems, 3D point cloud data generated by sonar and depth sensors facilitates the execution of more complex collaborative tasks among robots, which partly rely on efficient data transmission. In this work, we propose a hybrid encoder framework based on convolutional neural network and Transformer for underwater point cloud transmission, aimed at dealing with large-scale point cloud data. Our encoder allows setting lower compression rate on regions or objects of interest for semantic point clouds, preserving crucial information in the point cloud. For underwater acoustic communication, we employ orthogonal frequency division multiplexing combined with deep joint source-channel coding for transmission to enhance the system’s error-resilience. Compared to state-of-the-art methods in the simulation experiments, our end-to-end framework achieves considerable compression performance while eliminating certain cliff and leveling effects, also demonstrating robustness even with changing channels.
- 24 October 2024
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (1) , 015111
Underwater point cloud transmission framework: hybrid encoder implementation based on CNN and transformer
In underwater multi-robot systems, 3D point cloud data generated by sonar and depth sensors facilitates the execution of more complex collaborative tasks among robots, which partly rely on efficient data transmission. In this work, we propose a hybrid encoder framework based on convolutional neural network and Transformer for underwater point cloud transmission, aimed at dealing with large-scale point cloud data. Our encoder allows setting lower compression rate on regions or objects of interest for semantic point clouds, preserving crucial information in the point cloud. For underwater acoustic communication, we employ orthogonal frequency division multiplexing combined with deep joint source-channel coding for transmission to enhance the system’s error-resilience. Compared to state-of-the-art methods in the simulation experiments, our end-to-end framework achieves considerable compression performance while eliminating certain cliff and leveling effects, also demonstrating robustness even with changing channels.
- Open Access
- 9 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 149381-149390
Research on Unknown Space Target Pose Estimation Method Based on Point Cloud
As space in-orbit missions continue to develop, many uncooperative unknown space targets have occupied valuable orbital resources, seriously threatening the safety of spacecraft in orbit operation. One of the key technologies for in orbit space services is the determination of the relative attitude between the target and the tracker. The existing research mainly focuses on accurately determining the position and attitude of large satellites with known features such as solar panels, and lacks research verification for small space targets with complex shapes. This article first proposes a method for synthesizing visual perception data of space targets, synthesizing a batch of space targets perception data containing space fragments. In order to achieve accurate pose estimation of space targets with various shape features, a two-stage pose estimation method combining random sampling consistency algorithm and improved ICP algorithm is proposed. Finally, this article conducted extensive experimental validation on the proposed synthetic point cloud dataset. Compared with other point cloud registration algorithms, the proposed method achieves better registration results on synthesized multi scene simulation data.
- 9 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 149381-149390
Open AccessResearch on Unknown Space Target Pose Estimation Method Based on Point Cloud
As space in-orbit missions continue to develop, many uncooperative unknown space targets have occupied valuable orbital resources, seriously threatening the safety of spacecraft in orbit operation. One of the key technologies for in orbit space services is the determination of the relative attitude between the target and the tracker. The existing research mainly focuses on accurately determining the position and attitude of large satellites with known features such as solar panels, and lacks research verification for small space targets with complex shapes. This article first proposes a method for synthesizing visual perception data of space targets, synthesizing a batch of space targets perception data containing space fragments. In order to achieve accurate pose estimation of space targets with various shape features, a two-stage pose estimation method combining random sampling consistency algorithm and improved ICP algorithm is proposed. Finally, this article conducted extensive experimental validation on the proposed synthetic point cloud dataset. Compared with other point cloud registration algorithms, the proposed method achieves better registration results on synthesized multi scene simulation data.
- 5 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 31 (10709908) , 2475-2479
Unsupervised Low-Light Image Enhancement via Feature Smoothing and Curve Regression
Various methods have been discussed in the field of Low-light Image Enhancement (LIE), however the inability to effectively restrain noise still led to unsupervised LIE remaining a challenge. Traditional supervised LIE methods had the advantage of suppressing noise, but they suffer from generalization problems since the pairing of low-light and normal-light images are always required in the process. To this end, this work introduces an unsupervised LIE enhancement method that eliminates the need for paired images and effectively suppresses noise. The method utilizes the Retinex image decomposition technique to split the image into illumination and reflection components and processes the features of each part separately. In addition, we designed two deep learning architectures, lighting net (LINet) and denoising net (DNNet), to brighten the illumination component and denoise the reflection component respectively. We validate our method on two well-known open datasets and compared it with the state-of-the-art approaches. The experiments show that that the proposed method performs well in dimly lit scenes and visual verification, all quantitative metrics have at least 8% improvement on the best performing validation set.
- 5 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 31 (10709908) , 2475-2479
Unsupervised Low-Light Image Enhancement via Feature Smoothing and Curve Regression
Various methods have been discussed in the field of Low-light Image Enhancement (LIE), however the inability to effectively restrain noise still led to unsupervised LIE remaining a challenge. Traditional supervised LIE methods had the advantage of suppressing noise, but they suffer from generalization problems since the pairing of low-light and normal-light images are always required in the process. To this end, this work introduces an unsupervised LIE enhancement method that eliminates the need for paired images and effectively suppresses noise. The method utilizes the Retinex image decomposition technique to split the image into illumination and reflection components and processes the features of each part separately. In addition, we designed two deep learning architectures, lighting net (LINet) and denoising net (DNNet), to brighten the illumination component and denoise the reflection component respectively. We validate our method on two well-known open datasets and compared it with the state-of-the-art approaches. The experiments show that that the proposed method performs well in dimly lit scenes and visual verification, all quantitative metrics have at least 8% improvement on the best performing validation set.
- Open Access
- 18 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 107150-107159
Gaze Estimation Based on Attention Mechanism Combined With Temporal Network
Due to the continuous and dynamic nature of gaze estimation, the true gaze point at each moment is closely related to the previous moment. Simply detecting individual frames of facial images cannot yield accurate gaze information. In current CNN-based gaze estimation methods, the effective utilization of eye movement temporal information and the ability to capture global relationships in the feature extraction process remain problematic. Addressing these concerns, this paper proposes a novel gaze estimation framework, named FE-net, which incorporates a temporal network. This framework introduces channel attention modules and self-attention modules, enhancing the comprehensive utilization of extracted features and reinforcing the contribution of valuable regions to gaze estimation. We further integrate an RNN structure to learn the temporal dynamics of eye movement processes, significantly improving gaze direction prediction accuracy. This framework predicts the gaze directions of left and right eyes separately using monocular and facial features and computes the overall gaze direction. FE-net achieves state-of-the-art accuracy of 3.19° and 3.16° on the EVE dataset and the MPIIFaceGaze dataset, respectively.
- 18 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 107150-107159
Open AccessGaze Estimation Based on Attention Mechanism Combined With Temporal Network
Due to the continuous and dynamic nature of gaze estimation, the true gaze point at each moment is closely related to the previous moment. Simply detecting individual frames of facial images cannot yield accurate gaze information. In current CNN-based gaze estimation methods, the effective utilization of eye movement temporal information and the ability to capture global relationships in the feature extraction process remain problematic. Addressing these concerns, this paper proposes a novel gaze estimation framework, named FE-net, which incorporates a temporal network. This framework introduces channel attention modules and self-attention modules, enhancing the comprehensive utilization of extracted features and reinforcing the contribution of valuable regions to gaze estimation. We further integrate an RNN structure to learn the temporal dynamics of eye movement processes, significantly improving gaze direction prediction accuracy. This framework predicts the gaze directions of left and right eyes separately using monocular and facial features and computes the overall gaze direction. FE-net achieves state-of-the-art accuracy of 3.19° and 3.16° on the EVE dataset and the MPIIFaceGaze dataset, respectively.
- 1 June 2023
- Published by American Institute of Aeronautics and Astronautics (AIAA) in Journal of Aerospace Information Systems
- Vol. 20 (6) , 334-342
Monocular Depth Estimation of Noncooperative Spacecraft Based on Deep Learning
Estimating position and stereographic geometry information of noncooperative spacecraft through three-dimensional (3-D) reconstruction techniques is of great significance. Depth information acquisition is an important component of 3-D reconstruction. Monocular cameras are cheaper and more widely used than depth sensors. The relative positions of noncooperative spacecraft and our spacecraft can be calculated from depth maps of monocular depth estimation and the camera parameters, providing data support for subsequent tracking and capture missions. This paper proposes a monocular depth estimation network combining the convolutional neural network (CNN) and a vision transformer (VIT) to improve the prediction accuracy of few-shot samples. We extract detail features and global features from the CNN and VIT encoders, respectively, and then fuse deep features and shallow features by a skip-connected upsampling decoder. Compared with the representative depth estimation algorithms in recent years on the NYU-Depth V2 dataset, the proposed network structure combines the advantages of the CNN and VIT as well as estimates the global depth of the scene more accurately while maintaining details. To solve the lack of spacecraft data collection, a new dataset is made from 3-D simulation models. Experiments on the self-made dataset demonstrate the feasibility of this method in aerospace engineering.
- 1 June 2023
- Published by American Institute of Aeronautics and Astronautics (AIAA) in Journal of Aerospace Information Systems
- Vol. 20 (6) , 334-342
Monocular Depth Estimation of Noncooperative Spacecraft Based on Deep Learning
Estimating position and stereographic geometry information of noncooperative spacecraft through three-dimensional (3-D) reconstruction techniques is of great significance. Depth information acquisition is an important component of 3-D reconstruction. Monocular cameras are cheaper and more widely used than depth sensors. The relative positions of noncooperative spacecraft and our spacecraft can be calculated from depth maps of monocular depth estimation and the camera parameters, providing data support for subsequent tracking and capture missions. This paper proposes a monocular depth estimation network combining the convolutional neural network (CNN) and a vision transformer (VIT) to improve the prediction accuracy of few-shot samples. We extract detail features and global features from the CNN and VIT encoders, respectively, and then fuse deep features and shallow features by a skip-connected upsampling decoder. Compared with the representative depth estimation algorithms in recent years on the NYU-Depth V2 dataset, the proposed network structure combines the advantages of the CNN and VIT as well as estimates the global depth of the scene more accurately while maintaining details. To solve the lack of spacecraft data collection, a new dataset is made from 3-D simulation models. Experiments on the self-made dataset demonstrate the feasibility of this method in aerospace engineering.
- 26 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 30 (10709908) , 120-124
Two-Stage Underwater Image Restoration Algorithm Based on Physical Model and Causal Intervention
Underwater images often suffer from severe color degradation, haze and local blur, which are caused by the scattering and absorption effects of light in water. Firstly, to address the lack of annotations in underwater images, we propose a multi-degradation rate underwater image degradation model. Additionally, we use the generative adversarial network (GAN) to guide the underwater image degradation model and generate paired images that can be used for underwater image restoration (UIR) network training. Also, given the problems of poor restoration quality, serious loss of detail, and slow inference speed of existing deep learning algorithms, we construct an underwater image restoration network that can reference in real-time. Moreover, through causal interventions in the image generation process, spurious correlations between global features and detailed features are eliminated. As a result, the detail generation ability of the image is improved. Experiments on real underwater image datasets demonstrate that compared with existing methods, our proposed method more effectively solves the problem of underwater image degradation.
- 26 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 30 (10709908) , 120-124
Two-Stage Underwater Image Restoration Algorithm Based on Physical Model and Causal Intervention
Underwater images often suffer from severe color degradation, haze and local blur, which are caused by the scattering and absorption effects of light in water. Firstly, to address the lack of annotations in underwater images, we propose a multi-degradation rate underwater image degradation model. Additionally, we use the generative adversarial network (GAN) to guide the underwater image degradation model and generate paired images that can be used for underwater image restoration (UIR) network training. Also, given the problems of poor restoration quality, serious loss of detail, and slow inference speed of existing deep learning algorithms, we construct an underwater image restoration network that can reference in real-time. Moreover, through causal interventions in the image generation process, spurious correlations between global features and detailed features are eliminated. As a result, the detail generation ability of the image is improved. Experiments on real underwater image datasets demonstrate that compared with existing methods, our proposed method more effectively solves the problem of underwater image degradation.
- Open Access
- 19 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 132515-132524
Efficient Real-Time Tracking of Satellite Components Based on Frame Matching
In order to obtain the satellite’s in-orbit attitude information, it is necessary to track the satellite components in satellite video sequences. To solve the problem of low illumination and target occlusion in space environment, we propose an efficient satellite component tracking technique based on Rethinking Space-Time Networks with Improved Memory Coverage (STCN). We classify the pixels in the query frame by feature matching network that establishes the corresponding relationship between the frames. Unlike STCN, we reduce the contribution of background region in feature matching and enhance the robustness of the model in low illumination environment, thus improving the segmentation results. For lost targets due to the overturning and occlusion of satellite components, a position information encoder module is designed to further raise the tracking performance of the model. In addition, we present a local matching module to upgrade the existing feature matching methods. Experiments demonstrate that compared to STCN, our method heightens the tracking performance (J&F) by 10.1% and can achieve multi-object recognition at 15+ FPS.
- 19 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 132515-132524
Open AccessEfficient Real-Time Tracking of Satellite Components Based on Frame Matching
In order to obtain the satellite’s in-orbit attitude information, it is necessary to track the satellite components in satellite video sequences. To solve the problem of low illumination and target occlusion in space environment, we propose an efficient satellite component tracking technique based on Rethinking Space-Time Networks with Improved Memory Coverage (STCN). We classify the pixels in the query frame by feature matching network that establishes the corresponding relationship between the frames. Unlike STCN, we reduce the contribution of background region in feature matching and enhance the robustness of the model in low illumination environment, thus improving the segmentation results. For lost targets due to the overturning and occlusion of satellite components, a position information encoder module is designed to further raise the tracking performance of the model. In addition, we present a local matching module to upgrade the existing feature matching methods. Experiments demonstrate that compared to STCN, our method heightens the tracking performance (J&F) by 10.1% and can achieve multi-object recognition at 15+ FPS.