Xiaohong Yan
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
Xiaohong Yan has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Fengqi Li
Dalian Jiaotong University
6 shared publications - Fengqiang Xu
School of Railway Intelligent Engineering, Dalian Jiaotong University, Dalian, China
4 shared publications - Xianping Fu
Dalian Maritime University
3 shared publications Guangyuan Wang
Information Science and Technology School, Dalian Maritime University, Dalian, China
2 shared publications- Renteng Qu2 shared publications
- Yanjuan Wang
School of Railway Intelligent Engineering, Dalian Jiaotong University, Dalian 116028, China
2 shared publications
Latest Publications
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 51 (4) , 3737-3754
CHUIG: Controllable High-Resolution Underwater Image Generator With Refined Diffusion Model
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 51 (4) , 3737-3754
CHUIG: Controllable High-Resolution Underwater Image Generator With Refined Diffusion Model
- 8 September 2026
- Published by Elsevier in Computers & Graphics
- Vol. 140
WaveDiff: Wavelet-guided latent diffusion with Fourier detail enhancement for underwater images
- 8 September 2026
- Published by Elsevier in Computers & Graphics
- Vol. 140
WaveDiff: Wavelet-guided latent diffusion with Fourier detail enhancement for underwater images
- 19 June 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (25) , 256115
A robust rotating machinery fault diagnosis method under high noise based on hybrid multi-scale feature extraction and mamba-inspired linear transformer
Rotating machinery is widely used in industries such as equipment manufacturing and transportation. However, its long-term operation often leads to machinery faults, which can result in huge economic losses and safety risks. Therefore, reliable diagnostic methods are essential to ensure the safety and efficacy of systems. Recently, Transformer-based methods have excelled in fault diagnosis. Nevertheless, high noise interference can significantly degrade the performance of models, making robust diagnosis difficult. To address this issue, this paper proposes MDCambaFormer, a novel robust fault diagnosis method. First, a multi-scale deep dilated convolution module is proposed to effectively enhance the representational response of the model to genuine weak fault impulses in noisy environments. By introducing dilation factors into deep convolutions, this module leverages the semantic complementarity of features across multi-scale branches to suppress noise interference. Second, a Mamba-inspired gated linear attention (MGLA) mechanism is developed to dynamically select critical signal components and filter out noise via an adaptive gating design, utilizing a linear attention mechanism integrated with Mamba dual-branch blocks to perform global correlation modeling of key fault features, which significantly strengthens model robustness against high-frequency noise interference. Furthermore, a lightweight multi-layer perceptron embedded within the MGLA is employed to further reduce channel redundancy. Extensive experimental results on three datasets (BJTU-RAO, JNU, and Ottawa) show that under high noise interference, MDCambaFormer achieves higher accuracy and stronger robustness compared to state-of-the-art convolutional neural network and Transformer methods.
- 19 June 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (25) , 256115
A robust rotating machinery fault diagnosis method under high noise based on hybrid multi-scale feature extraction and mamba-inspired linear transformer
Rotating machinery is widely used in industries such as equipment manufacturing and transportation. However, its long-term operation often leads to machinery faults, which can result in huge economic losses and safety risks. Therefore, reliable diagnostic methods are essential to ensure the safety and efficacy of systems. Recently, Transformer-based methods have excelled in fault diagnosis. Nevertheless, high noise interference can significantly degrade the performance of models, making robust diagnosis difficult. To address this issue, this paper proposes MDCambaFormer, a novel robust fault diagnosis method. First, a multi-scale deep dilated convolution module is proposed to effectively enhance the representational response of the model to genuine weak fault impulses in noisy environments. By introducing dilation factors into deep convolutions, this module leverages the semantic complementarity of features across multi-scale branches to suppress noise interference. Second, a Mamba-inspired gated linear attention (MGLA) mechanism is developed to dynamically select critical signal components and filter out noise via an adaptive gating design, utilizing a linear attention mechanism integrated with Mamba dual-branch blocks to perform global correlation modeling of key fault features, which significantly strengthens model robustness against high-frequency noise interference. Furthermore, a lightweight multi-layer perceptron embedded within the MGLA is employed to further reduce channel redundancy. Extensive experimental results on three datasets (BJTU-RAO, JNU, and Ottawa) show that under high noise interference, MDCambaFormer achieves higher accuracy and stronger robustness compared to state-of-the-art convolutional neural network and Transformer methods.
- 27 April 2026
- Published by Elsevier in Infrared Physics & Technology
- Vol. 156
TriFusion: Three-stage image fusion paradigm with dual supervision learning for multi-modality image fusion
- 27 April 2026
- Published by Elsevier in Infrared Physics & Technology
- Vol. 156
TriFusion: Three-stage image fusion paradigm with dual supervision learning for multi-modality image fusion
- 7 August 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 297
A multi-scale feature extraction and attention aggregation network for underwater image enhancement
- 7 August 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 297
A multi-scale feature extraction and attention aggregation network for underwater image enhancement
- 15 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 50 (1) , 150-163
CFPNet: Complementary Feature Perception Network for Underwater Image Enhancement
Images shot underwater are usually characterized by global nonuniform information loss due to selective light absorption and scattering, resulting in various degradation problems, such as color distortion and low visibility. Recently, deep learning has drawn much attention in the field of underwater image enhancement (UIE) for its powerful performance. However, most deep learning-based UIE models rely on either pure convolutional neural network (CNN) or pure transformer, which makes it challenging to enhance images while maintaining local representations and global features simultaneously. In this article, we propose a novel complementary feature perception network (CFPNet), which embeds the transformer into the classical CNN-based UNet3+. The core idea is to fuse the advantages of CNN and transformer to obtain satisfactory high-quality underwater images that can naturally perceive local and global features. CFPNet employs a novel dual encoder structure of the CNN and transformer in parallel, while the decoder is composed of one trunk decoder and two auxiliary decoders. First, we propose the regionalized two-stage vision transformer that can progressively eliminate the variable levels of degradation in a coarse-to-fine manner. Second, we design the full-scale feature fusion module to explore sufficient information by merging the multiscale features. In addition, we propose an auxiliary feature guided learning strategy that utilizes reflectance and shading maps to guide the generation of the final results. The advantage of this strategy is to avoid repetitive and ineffective learning of the model, and to accomplish color correction and deblurring tasks more efficiently. Experiments demonstrate that our CFPNet can obtain high-quality underwater images and show superior performance compared to the state-of-the-art UIE methods qualitatively and quantitatively.
- 15 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Oceanic Engineering
- Vol. 50 (1) , 150-163
CFPNet: Complementary Feature Perception Network for Underwater Image Enhancement
Images shot underwater are usually characterized by global nonuniform information loss due to selective light absorption and scattering, resulting in various degradation problems, such as color distortion and low visibility. Recently, deep learning has drawn much attention in the field of underwater image enhancement (UIE) for its powerful performance. However, most deep learning-based UIE models rely on either pure convolutional neural network (CNN) or pure transformer, which makes it challenging to enhance images while maintaining local representations and global features simultaneously. In this article, we propose a novel complementary feature perception network (CFPNet), which embeds the transformer into the classical CNN-based UNet3+. The core idea is to fuse the advantages of CNN and transformer to obtain satisfactory high-quality underwater images that can naturally perceive local and global features. CFPNet employs a novel dual encoder structure of the CNN and transformer in parallel, while the decoder is composed of one trunk decoder and two auxiliary decoders. First, we propose the regionalized two-stage vision transformer that can progressively eliminate the variable levels of degradation in a coarse-to-fine manner. Second, we design the full-scale feature fusion module to explore sufficient information by merging the multiscale features. In addition, we propose an auxiliary feature guided learning strategy that utilizes reflectance and shading maps to guide the generation of the final results. The advantage of this strategy is to avoid repetitive and ineffective learning of the model, and to accomplish color correction and deblurring tasks more efficiently. Experiments demonstrate that our CFPNet can obtain high-quality underwater images and show superior performance compared to the state-of-the-art UIE methods qualitatively and quantitatively.
- 21 August 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83 (9) , 25519-25542
An image quality-aware approach with adaptive scattering coefficients for single image dehazing
- 21 August 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83 (9) , 25519-25542
An image quality-aware approach with adaptive scattering coefficients for single image dehazing
- 31 July 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83 (7) , 20169-20192
Underwater image dehazing using a novel color channel based dual transmission map estimation
- 31 July 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83 (7) , 20169-20192
Underwater image dehazing using a novel color channel based dual transmission map estimation
- 27 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3, 996-1002
5G New Radio with Orthogonal Multiwavelet Packet Division Multiplexing
A novel 5G waveform based with orthogonal multiwavelet packet was proposed in this paper. Fast Fourier transform has been adopted in 5G. The 5G symbol overlaps only in frequency domain and which leads to a large subcarrier number. But too many subcarriers will lead to a high peak to average power ratio (P APR). Due to the multiwavelet packet's orthogonal characteristics in frequency domain and integral shift in time domain, the proposed multiwavelet packet division multiplexing(MWPTDM) 5G NR symbol not only overlaps in frequency domain, but also overlaps with integral time shift in time domain. Thus the sub carrier number can be decreased significantly and the PAPR is much lower than FFT-OFDM. The MWPTDM also has higher band efficiency and lower sidelobes than FFT-OFDM and wavelet packet OFDM (WPT-OFDM), and lower computational complexity than FFT-OFDM 5G NR.
- 27 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3, 996-1002
5G New Radio with Orthogonal Multiwavelet Packet Division Multiplexing
A novel 5G waveform based with orthogonal multiwavelet packet was proposed in this paper. Fast Fourier transform has been adopted in 5G. The 5G symbol overlaps only in frequency domain and which leads to a large subcarrier number. But too many subcarriers will lead to a high peak to average power ratio (P APR). Due to the multiwavelet packet's orthogonal characteristics in frequency domain and integral shift in time domain, the proposed multiwavelet packet division multiplexing(MWPTDM) 5G NR symbol not only overlaps in frequency domain, but also overlaps with integral time shift in time domain. Thus the sub carrier number can be decreased significantly and the PAPR is much lower than FFT-OFDM. The MWPTDM also has higher band efficiency and lower sidelobes than FFT-OFDM and wavelet packet OFDM (WPT-OFDM), and lower computational complexity than FFT-OFDM 5G NR.
- 22 February 2022
- Published by Elsevier in Signal Processing: Image Communication
- Vol. 104
A novel biologically-inspired method for underwater image enhancement
- 22 February 2022
- Published by Elsevier in Signal Processing: Image Communication
- Vol. 104
A novel biologically-inspired method for underwater image enhancement