Xiaotao Shao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Xiaotao Shao has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Zhongli Wang
School of Automation and Intelligence, Beijing Jiaotong University, Beijing, China
10 shared publications - Yan Shen
School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, China
9 shared publications - Wei Yang
Beijing Jiaotong University
4 shared publications - Tengfei Lu
School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, China
3 shared publications - Donglin Zhu
Beijing Jiaotong University, Beijing, China
2 shared publications - Wenlong Tian
School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing 100044, China
2 shared publications
Latest Publications
- 11 October 2026
- Published by Elsevier in Information Fusion
- Vol. 139
USSI-Net: Physics-guided unified spectral–spatial information fusion network for hyperspectral snapshot reconstruction
- 11 October 2026
- Published by Elsevier in Information Fusion
- Vol. 139
USSI-Net: Physics-guided unified spectral–spatial information fusion network for hyperspectral snapshot reconstruction
- 23 October 2025
- Published by Elsevier in Measurement
- Vol. 258
OrthSynergy: Orthogonal feature-based synergy descriptor for loop closure detection in large-scale scenes
- 23 October 2025
- Published by Elsevier in Measurement
- Vol. 258
OrthSynergy: Orthogonal feature-based synergy descriptor for loop closure detection in large-scale scenes
- 1 September 2025
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 84 (41) , 49397-49418
GLPF-Net:Global and Local-Physics-aware fusion network for single image dehazing
- 1 September 2025
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 84 (41) , 49397-49418
GLPF-Net:Global and Local-Physics-aware fusion network for single image dehazing
- 25 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 1786-1799
MHTraj: A Multi-Domain Hybrid Graph Neural Network With Causal-Spatial Modeling for Multi-Agent Trajectory Prediction
As a crucial part of social behavior modeling, multi-agent trajectory prediction with dynamic interactions has become one of the core technologies for autonomous systems. One of the main research streams is to leverage relational graph Transformers to learn interactions among agents, as this method demonstrates superior capability in capturing implicit features and modeling complex relationships. However, such researches fall short in efficiently utilizing multi-domain features, consequently being constrained by spatial attention layers that hinder accuracy improvements. To address the above limitations and achieve a more lightweight map-free architecture, this work introduces a Multi-domain Hybrid Graph Transformer (MHTraj) for predicting multi-modal future trajectories. The core insight of the method lies in its well-designed encoder, which enhances inputs via a frequency-domain mixer and employs causal convolution attention to capture global temporal information. Furthermore, a graph Transformer is incorporated to aggregate spatial relationships, thereby yielding the final hybrid features. Extensive experiments on 4 real-world datasets (NBA, ETH/UCY, SDD, nuScenes) show that the proposed method achieves competitive state-of-the-art (SOTA) performances, and meanwhile reducing the parameter count by nearly 50% on the ETH/UCY datasets.
- 25 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 1786-1799
MHTraj: A Multi-Domain Hybrid Graph Neural Network With Causal-Spatial Modeling for Multi-Agent Trajectory Prediction
As a crucial part of social behavior modeling, multi-agent trajectory prediction with dynamic interactions has become one of the core technologies for autonomous systems. One of the main research streams is to leverage relational graph Transformers to learn interactions among agents, as this method demonstrates superior capability in capturing implicit features and modeling complex relationships. However, such researches fall short in efficiently utilizing multi-domain features, consequently being constrained by spatial attention layers that hinder accuracy improvements. To address the above limitations and achieve a more lightweight map-free architecture, this work introduces a Multi-domain Hybrid Graph Transformer (MHTraj) for predicting multi-modal future trajectories. The core insight of the method lies in its well-designed encoder, which enhances inputs via a frequency-domain mixer and employs causal convolution attention to capture global temporal information. Furthermore, a graph Transformer is incorporated to aggregate spatial relationships, thereby yielding the final hybrid features. Extensive experiments on 4 real-world datasets (NBA, ETH/UCY, SDD, nuScenes) show that the proposed method achieves competitive state-of-the-art (SOTA) performances, and meanwhile reducing the parameter count by nearly 50% on the ETH/UCY datasets.
- 4 July 2025
- Published by Elsevier in Computer Vision and Image Understanding
- Vol. 259
AdaptDiff: Adaptive diffusion learning for low-light image enhancement
- 4 July 2025
- Published by Elsevier in Computer Vision and Image Understanding
- Vol. 259
AdaptDiff: Adaptive diffusion learning for low-light image enhancement
- 29 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (7) , 7484-7491
Hybrid-NeuroSLAM: A Neurobiologically Inspired Hybrid Visual-Inertial SLAM Method for Large Scale Environment
Animals in nature exhibit remarkable spatial cognition abilities, enabling them to achieve long-distance autonomous navigation efficiently in unknown environments. Neurobiologically inspired SLAM methods primarily aim to mimic this capability of spatial cognition and navigation. However, existing works have only been able to perform rough velocity estimation with scanline intensity profiles, which may lead to a severely distorted cognitive map, particularly in large scale environment. This work presents a novel neurobiologically inspired Visual-Inertial SLAM method, called Hybrid-NueroSLAM, that addresses two key issues: precisely and robust visual-inertial fusion-based motion estimation as input for cognition map, and the map consistency for large scale environment. Based on high-precision Visual-Inertial odometry, a precise and robust visual memory mechanism, and a neural cell network based on our proposed improved Bayesian attractor for spatial cognition, it constructs a semi-metric topological experience map. The experimental results with benchmark datasets KITTI and Complex Urban, demonstrate that the proposed method effectively mitigates map jitters and maintains high accuracy and robustness, even on a 11 km urban dataset. Source code: https://github.com/iRobotVision/Hybird-NeuroSLAM
- 29 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (7) , 7484-7491
Hybrid-NeuroSLAM: A Neurobiologically Inspired Hybrid Visual-Inertial SLAM Method for Large Scale Environment
Animals in nature exhibit remarkable spatial cognition abilities, enabling them to achieve long-distance autonomous navigation efficiently in unknown environments. Neurobiologically inspired SLAM methods primarily aim to mimic this capability of spatial cognition and navigation. However, existing works have only been able to perform rough velocity estimation with scanline intensity profiles, which may lead to a severely distorted cognitive map, particularly in large scale environment. This work presents a novel neurobiologically inspired Visual-Inertial SLAM method, called Hybrid-NueroSLAM, that addresses two key issues: precisely and robust visual-inertial fusion-based motion estimation as input for cognition map, and the map consistency for large scale environment. Based on high-precision Visual-Inertial odometry, a precise and robust visual memory mechanism, and a neural cell network based on our proposed improved Bayesian attractor for spatial cognition, it constructs a semi-metric topological experience map. The experimental results with benchmark datasets KITTI and Complex Urban, demonstrate that the proposed method effectively mitigates map jitters and maintains high accuracy and robustness, even on a 11 km urban dataset. Source code: https://github.com/iRobotVision/Hybird-NeuroSLAM
- Open Access
- 1 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 178-182
Non-homogeneous Haze Removal Based on Attentional Feature Enhancement in Encoder-decoder Networks
Existing image dehazing methods have low quality in dealing with non-uniform haze in real scenes, especially in heavy haze situations. To address this problem, we propose a Multi-Level Feature Extraction Improvement Network (MLFEIN), which effectively removes non-uniform haze in images by extracting and fusing features at different stages and levels. Specifically, we design an Efficient Feature Fusion (EFF) which has an encoder-decoder structure with fusion mechanism, and a Channel and Spatial Information Enhancement (CSIE) module that can efficiently extract features. The EFF module can preserve the image structure information and eliminate artifacts caused by non-uniform haze by combining high-level and low-level features. The CSIE module can further eliminate non-uniform haze by using two separate attention blocks to generate information maps, which utilize spatial and channel information respectively. Moreover, we add residual structures before the attention mechanism to prevent network degradation and enhance global feature extraction. Extensive experiments show that our method achieves better results than the state-of-the-art methods.
- 1 March 2024
- Published by Association for Computing Machinery (ACM)
- p. 178-182
Open AccessNon-homogeneous Haze Removal Based on Attentional Feature Enhancement in Encoder-decoder Networks
Existing image dehazing methods have low quality in dealing with non-uniform haze in real scenes, especially in heavy haze situations. To address this problem, we propose a Multi-Level Feature Extraction Improvement Network (MLFEIN), which effectively removes non-uniform haze in images by extracting and fusing features at different stages and levels. Specifically, we design an Efficient Feature Fusion (EFF) which has an encoder-decoder structure with fusion mechanism, and a Channel and Spatial Information Enhancement (CSIE) module that can efficiently extract features. The EFF module can preserve the image structure information and eliminate artifacts caused by non-uniform haze by combining high-level and low-level features. The CSIE module can further eliminate non-uniform haze by using two separate attention blocks to generate information maps, which utilize spatial and channel information respectively. Moreover, we add residual structures before the attention mechanism to prevent network degradation and enhance global feature extraction. Extensive experiments show that our method achieves better results than the state-of-the-art methods.
- 20 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-12
DefVAE: A Defect Detection Method for Catenary Devices Based on Variational Autoencoder
Catenary is one of the most crucial parts of electrified railway system. How to detect the part defects of catenary in time and keep it in a stable and safe operation conditions are the main tasks for maintainance. The existing data-driven vision-based defect detection methods will face a big challenge that, there are many catenary parts and each part has several types of defects, but there are few defect sample images for each type of defect, which limits them in real applications. To alleviate this problem, a semantic label-enhanced variational autoencoder (VAE) method for catenary part defect detection, termed Defect VAE (DefVAE), is presented. The proposed method is based on a label-enhanced VAE network that determines the distribution boundary in latent feature space of defect sample for additional defect generation. The addition of semantic label information to the VAE improves the inter-class distance in latent space, clarifying the boundary and further boosting the capabilities of the defect detection method, which is demonstrated in our experiments. Additionally, the defect type is classified by combining the outputs of the classifier confidence and the pixel-level reconstruction error which is based on a sliding label mode of the variational autoencoder. Extensive experiments with the open benchmark dataset MVTec and the catenary dataset collected by ourselves demonstrate that the presented DefVAE outperforms the baseline methods across the majority of indicators.
- 20 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-12
DefVAE: A Defect Detection Method for Catenary Devices Based on Variational Autoencoder
Catenary is one of the most crucial parts of electrified railway system. How to detect the part defects of catenary in time and keep it in a stable and safe operation conditions are the main tasks for maintainance. The existing data-driven vision-based defect detection methods will face a big challenge that, there are many catenary parts and each part has several types of defects, but there are few defect sample images for each type of defect, which limits them in real applications. To alleviate this problem, a semantic label-enhanced variational autoencoder (VAE) method for catenary part defect detection, termed Defect VAE (DefVAE), is presented. The proposed method is based on a label-enhanced VAE network that determines the distribution boundary in latent feature space of defect sample for additional defect generation. The addition of semantic label information to the VAE improves the inter-class distance in latent space, clarifying the boundary and further boosting the capabilities of the defect detection method, which is demonstrated in our experiments. Additionally, the defect type is classified by combining the outputs of the classifier confidence and the pixel-level reconstruction error which is based on a sliding label mode of the variational autoencoder. Extensive experiments with the open benchmark dataset MVTec and the catenary dataset collected by ourselves demonstrate that the presented DefVAE outperforms the baseline methods across the majority of indicators.
- Open Access
- 7 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 162325-162336
A High-Precision Dual-Frequency Magnetic Induction Through-the-Earth Positioning Method
To achieve the high-precision through-the-earth positioning in mine rescue, we propose a dual-frequency magnetic induction positioning method. The dual-frequency transmitter underground is a loop energized coil. The receiver on the ground is a three-axis magnetic core coil. We first carry on the algebraic transformation to the expression of the dual-frequency positioning signal, and obtain the transceiver distance function that is not affected by the layered earth conductivity. Then, through the tilt angle of the receiver and the rotation matrices, we correct the receiver tilt. By the direction of the geomagnetic field, we correct the direction deviation of the receiver. Finally, we obtain the three-dimensional coordinates of the transmitter by decomposition of the transceiver distance. The analysis and simulation results are as follows. The dual-frequency positioning method eliminates the positioning error caused by the path loss of the stratified earth media. The rotation matrices correct the error caused by the receiver tilt. When the receiver is close to the top of the transmitter, the positioning error can be reduced. By configuring proper transmitter parameters, we can maintain a balance between transmitter size and rescue efficiency.
- 7 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 162325-162336
Open AccessA High-Precision Dual-Frequency Magnetic Induction Through-the-Earth Positioning Method
To achieve the high-precision through-the-earth positioning in mine rescue, we propose a dual-frequency magnetic induction positioning method. The dual-frequency transmitter underground is a loop energized coil. The receiver on the ground is a three-axis magnetic core coil. We first carry on the algebraic transformation to the expression of the dual-frequency positioning signal, and obtain the transceiver distance function that is not affected by the layered earth conductivity. Then, through the tilt angle of the receiver and the rotation matrices, we correct the receiver tilt. By the direction of the geomagnetic field, we correct the direction deviation of the receiver. Finally, we obtain the three-dimensional coordinates of the transmitter by decomposition of the transceiver distance. The analysis and simulation results are as follows. The dual-frequency positioning method eliminates the positioning error caused by the path loss of the stratified earth media. The rotation matrices correct the error caused by the receiver tilt. When the receiver is close to the top of the transmitter, the positioning error can be reduced. By configuring proper transmitter parameters, we can maintain a balance between transmitter size and rescue efficiency.
- Open Access
- 17 June 2021
- Published by Institution of Engineering and Technology (IET) in IET Signal Processing
- Vol. 15 (9) , 573-583
Shearlet‐based compressed sensing with non‐local similarity for MRI breast image reconstruction
Magnetic resonance imaging (MRI) requires long detection time and makes patients uncomfortable. The proposed compressed sensing MRI compressed sensing with shearlet dictionary and non‐local similarity model is established with shearlet dictionary and non‐local similarity. The shearlet dictionary is adopted in MRI compressed sensing to represent breast tissues with sparser data in different scales and directions. The non‐local similarity of an image is integrated to the model to preserve the lesion details of the reconstructed MRI images. The proposed model is solved by the split Bregman algorithm to obtain the optimized image iteratively. Experiments are performed on practical MRI breast images with sampling data of 13% and 10%. With the decrease of sampling data, the proposed method can reconstruct the image with better visual effect and higher peak signal‐to‐noise ratio (PSNR) and structural similarity index (SSIM) than traditional methods. There is an improvement of 13 dB of PSNR and 0.2 of SSIM under 10% data. The proposed method can reconstruct MRI images with less data and higher reconstruction quality compared with the traditional methods.
- 17 June 2021
- Published by Institution of Engineering and Technology (IET) in IET Signal Processing
- Vol. 15 (9) , 573-583
Open AccessShearlet‐based compressed sensing with non‐local similarity for MRI breast image reconstruction
Magnetic resonance imaging (MRI) requires long detection time and makes patients uncomfortable. The proposed compressed sensing MRI compressed sensing with shearlet dictionary and non‐local similarity model is established with shearlet dictionary and non‐local similarity. The shearlet dictionary is adopted in MRI compressed sensing to represent breast tissues with sparser data in different scales and directions. The non‐local similarity of an image is integrated to the model to preserve the lesion details of the reconstructed MRI images. The proposed model is solved by the split Bregman algorithm to obtain the optimized image iteratively. Experiments are performed on practical MRI breast images with sampling data of 13% and 10%. With the decrease of sampling data, the proposed method can reconstruct the image with better visual effect and higher peak signal‐to‐noise ratio (PSNR) and structural similarity index (SSIM) than traditional methods. There is an improvement of 13 dB of PSNR and 0.2 of SSIM under 10% data. The proposed method can reconstruct MRI images with less data and higher reconstruction quality compared with the traditional methods.