Xiaofei 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
Xiaofei Zhang has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Chee-Meng Chew
Department of Mechanical Engineering, National University of Singapore, Queenstown, Singapore
3 shared publications - Guangming Li
National Innovation Institute of Defense Technology, Academy of Military Sciences, Beijing 100071, China
3 shared publications - Haocai Huang
Zhejiang University
3 shared publications - Jianfeng Li
College of Electronic and Information Engineering and the Key Laboratory of Dynamic Cognitive System of Electromagnetic Spectrum Space, Ministry of Industry and Information Technology, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu, China
3 shared publications Xinyi Xie
School of Intelligent Manufacturing Ecosystem, Xian Jiaotong-Liverpool University, Suzhou, China
3 shared publications- Desheng Weng
Sun Yat-sen University Cancer Center 1 Melanoma and Sarcoma Oncology Unit, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine Guangzhou, China
2 shared publications
Latest Publications
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4206513
Physics-Constrained Deep Neural Networks for Short-Term 3-D Flow Field Prediction Using Coastal Acoustic Tomography
Accurate prediction of three-dimensional (3-D) ocean flow fields is essential for understanding small-scale dynamic processes, improving environmental monitoring, and enhancing the performance of underwater observation and operation sys-tems. However, predictive modeling using coastal acoustic to-mography (CAT) data remains relatively unexplored. This study proposes a novel physics-constrained deep learning framework, termed PC-LSTM-MLP, to achieve the short-term prediction of 3-D ocean flow fields from CAT observations. The framework consists of three integrated components: a long short-term memory (LSTM) network for capturing temporal dependencies in range-averaged currents (RACs), a multilayer perceptron (MLP) for nonlinear mapping to spatial Fourier coefficients, and a physics-constrained (PC) module that imposes a horizontal incompressibility constraint to enhance the accuracy of the predicted 3-D flow fields. Experimental results on real CAT data demonstrate that the proposed PC-LSTM-MLP framework effec-tively captures complex spatiotemporal dynamics and provides accurate predictions up to 25 minutes ahead. The model achieves an average mean absolute error (MAE) of 0.0241 m/s and an average root mean square error (RMSE) of 0.0269 m/s in 3-D flow field prediction. This research shifts CAT analysis from traditional inversion methods to intelligent, prediction-oriented modeling, underscoring the significant potential of deep learning in ocean flow field prediction.
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4206513
Physics-Constrained Deep Neural Networks for Short-Term 3-D Flow Field Prediction Using Coastal Acoustic Tomography
Accurate prediction of three-dimensional (3-D) ocean flow fields is essential for understanding small-scale dynamic processes, improving environmental monitoring, and enhancing the performance of underwater observation and operation sys-tems. However, predictive modeling using coastal acoustic to-mography (CAT) data remains relatively unexplored. This study proposes a novel physics-constrained deep learning framework, termed PC-LSTM-MLP, to achieve the short-term prediction of 3-D ocean flow fields from CAT observations. The framework consists of three integrated components: a long short-term memory (LSTM) network for capturing temporal dependencies in range-averaged currents (RACs), a multilayer perceptron (MLP) for nonlinear mapping to spatial Fourier coefficients, and a physics-constrained (PC) module that imposes a horizontal incompressibility constraint to enhance the accuracy of the predicted 3-D flow fields. Experimental results on real CAT data demonstrate that the proposed PC-LSTM-MLP framework effec-tively captures complex spatiotemporal dynamics and provides accurate predictions up to 25 minutes ahead. The model achieves an average mean absolute error (MAE) of 0.0241 m/s and an average root mean square error (RMSE) of 0.0269 m/s in 3-D flow field prediction. This research shifts CAT analysis from traditional inversion methods to intelligent, prediction-oriented modeling, underscoring the significant potential of deep learning in ocean flow field prediction.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-15
Physics-Constrained Multilayer Perceptron for Flow Field Inversion in Coastal Acoustic Tomography
Accurate flow field inversion in coastal acoustic tomography (CAT) is crucial for improving ocean environmental monitoring, ecological management, marine engineering safety, and disaster early warning. The conventional tapered least squares (TLS) method primarily relies on Fourier expansion to represent the flow field and employs the L-curve method for regularization. However, this method often suffers from severe ill-posedness, leading to inaccurate inversion, local inversion anomalies, or even complete failure. To overcome these limitations, a novel inversion framework termed the physics-constrained multilayer perceptron (PCMLP) is proposed, which integrates physical constraints into a multilayer perceptron (MLP) architecture. Specifically, the MLP maps the forward-calculated range-averaged currents to the Fourier coefficients of the stream function, and a soft constraint penalizing the nonzero net flux through the domain boundary is incorporated to enhance the stability and accuracy of the inversion. Experimental results demonstrate that, when validated against acoustic Doppler current profiler (ADCP) measurements, the PCMLP model with λ = 0.01 reduces the mean absolute error of flow velocity inversion by 39.79% and the root mean square error by 33.01%, relative to the conventional TLS method, significantly improving the accuracy and stability of horizontal flow field inversion. After training, the PCMLP model requires approximately 19 times fewer multiply-accumulate operations per inference step than the TLS method, making it suitable for real-time or repeated deployments. The proposed method shows strong potential for broad applications in CAT-based flow field inversion and prediction, offering a robust and high-precision solution for ocean flow field observation.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-15
Physics-Constrained Multilayer Perceptron for Flow Field Inversion in Coastal Acoustic Tomography
Accurate flow field inversion in coastal acoustic tomography (CAT) is crucial for improving ocean environmental monitoring, ecological management, marine engineering safety, and disaster early warning. The conventional tapered least squares (TLS) method primarily relies on Fourier expansion to represent the flow field and employs the L-curve method for regularization. However, this method often suffers from severe ill-posedness, leading to inaccurate inversion, local inversion anomalies, or even complete failure. To overcome these limitations, a novel inversion framework termed the physics-constrained multilayer perceptron (PCMLP) is proposed, which integrates physical constraints into a multilayer perceptron (MLP) architecture. Specifically, the MLP maps the forward-calculated range-averaged currents to the Fourier coefficients of the stream function, and a soft constraint penalizing the nonzero net flux through the domain boundary is incorporated to enhance the stability and accuracy of the inversion. Experimental results demonstrate that, when validated against acoustic Doppler current profiler (ADCP) measurements, the PCMLP model with λ = 0.01 reduces the mean absolute error of flow velocity inversion by 39.79% and the root mean square error by 33.01%, relative to the conventional TLS method, significantly improving the accuracy and stability of horizontal flow field inversion. After training, the PCMLP model requires approximately 19 times fewer multiply-accumulate operations per inference step than the TLS method, making it suitable for real-time or repeated deployments. The proposed method shows strong potential for broad applications in CAT-based flow field inversion and prediction, offering a robust and high-precision solution for ocean flow field observation.
- 26 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. PP (01962892) , 1
Three-dimensional Flow Mapping for Monitoring Underwater Small-scale Dynamics Using Coastal Acoustic Tomography
In this study, coastal acoustic tomography (CAT) technology was utilized to perform an initial investigation into the observation of underwater small-scale dynamic processes. An underwater monitoring experiment using a CAT network was conducted with four stations at Huangcai Reservoir in Hunan, China. Ray acoustic simulations were employed to match the simulated ray paths of the vertical profile of each pair of trans-ducers to the two identified times of flight (TOFs). To visualize flow field fluctuations caused by the underwater small-scale dynamic processes, an autonomous underwater vehicle (AUV) was deployed to navigate within the observation area. Three-dimensional (3D) inversion techniques were used for multi-layer analyses of the underwater horizontal flow field in the region. The AUV cruising layer’s flow field was mapped and showed consistency with acoustic Doppler current profiler (ADCP) data. The velocity data recorded by the ADCP and CAT were 0.87 m/s and 0.82 m/s, respectively, as the AUV passed by. Additionally, the results were validated through comparison of range-average current measurements and the net inflow error. This study illustrates that the 3D CAT mapping method can effectively reconstruct the 3D flow field, facilitating the observation of underwater small-scale dynamic processes.
- 26 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. PP (01962892) , 1
Three-dimensional Flow Mapping for Monitoring Underwater Small-scale Dynamics Using Coastal Acoustic Tomography
In this study, coastal acoustic tomography (CAT) technology was utilized to perform an initial investigation into the observation of underwater small-scale dynamic processes. An underwater monitoring experiment using a CAT network was conducted with four stations at Huangcai Reservoir in Hunan, China. Ray acoustic simulations were employed to match the simulated ray paths of the vertical profile of each pair of trans-ducers to the two identified times of flight (TOFs). To visualize flow field fluctuations caused by the underwater small-scale dynamic processes, an autonomous underwater vehicle (AUV) was deployed to navigate within the observation area. Three-dimensional (3D) inversion techniques were used for multi-layer analyses of the underwater horizontal flow field in the region. The AUV cruising layer’s flow field was mapped and showed consistency with acoustic Doppler current profiler (ADCP) data. The velocity data recorded by the ADCP and CAT were 0.87 m/s and 0.82 m/s, respectively, as the AUV passed by. Additionally, the results were validated through comparison of range-average current measurements and the net inflow error. This study illustrates that the 3D CAT mapping method can effectively reconstruct the 3D flow field, facilitating the observation of underwater small-scale dynamic processes.
- 1 November 2019
- Published by Elsevier in Industrial Crops and Products
- Vol. 139, 111516
A novel application of the vacuum distillation technology in extracting Origanum vulgare L. essential oils
- 1 November 2019
- Published by Elsevier in Industrial Crops and Products
- Vol. 139, 111516
A novel application of the vacuum distillation technology in extracting Origanum vulgare L. essential oils
- 1 June 2018
- Published by Elsevier in Aerospace Science and Technology
- Vol. 77, 626-637
Starting characteristics and phenomenon of a supersonic wind tunnel coupled with inlet model
- 1 June 2018
- Published by Elsevier in Aerospace Science and Technology
- Vol. 77, 626-637
Starting characteristics and phenomenon of a supersonic wind tunnel coupled with inlet model
- Open Access
- 17 August 2017
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 11 (10) , 1581-1588
Two‐dimensional direction of arrival estimation for coprime planar arrays via a computationally efficient one‐dimensional partial spectral search approach
The authors investigate the problem of two‐dimensional (2D) direction of arrival (DOA) estimation of multiple signals for coprime planar arrays (CPAs) in this study and they propose a computationally efficient 1D partial spectral search approach based on multiple signal classification (MUSIC) algorithm. The conventional 2D MUSIC algorithm for CPAs has a great DOA estimation performance, but suffers from a tremendously expensive computational complexity due to the 2D spectral search. To this end, the proposed approach first decreases the dimension of the spectrum function to one dimension and then utilises the linear relationship between the true and ambiguous DOA estimates to form a 1D partial spectral search over a small sector. Finally, the true DOA estimates can be achieved based on the coprime property. The proposed approach can have an impressively good DOA estimation performance, but with a low computational cost. Simulation results validate the effectiveness and superiority of the proposed approach.
- 17 August 2017
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 11 (10) , 1581-1588
Open AccessTwo‐dimensional direction of arrival estimation for coprime planar arrays via a computationally efficient one‐dimensional partial spectral search approach
The authors investigate the problem of two‐dimensional (2D) direction of arrival (DOA) estimation of multiple signals for coprime planar arrays (CPAs) in this study and they propose a computationally efficient 1D partial spectral search approach based on multiple signal classification (MUSIC) algorithm. The conventional 2D MUSIC algorithm for CPAs has a great DOA estimation performance, but suffers from a tremendously expensive computational complexity due to the 2D spectral search. To this end, the proposed approach first decreases the dimension of the spectrum function to one dimension and then utilises the linear relationship between the true and ambiguous DOA estimates to form a 1D partial spectral search over a small sector. Finally, the true DOA estimates can be achieved based on the coprime property. The proposed approach can have an impressively good DOA estimation performance, but with a low computational cost. Simulation results validate the effectiveness and superiority of the proposed approach.
- Open Access
- 12 June 2017
- Published by Springer Nature in BMC Cancer
- Vol. 17 (1) , 1-10
Pooled safety analyses of ALK-TKI inhibitor in ALK-positive NSCLC
The anaplastic lymphoma kinase tyrosine kinase inhibitors (ALK-TKIs) have been administered to patients with ALK-positive non-small cell lung cancer for a long period of time and show a promising response. However, the differences in the toxicity profiles among these drugs are still unclear. We performed a comprehensive search of the MEDLINE, EMBASE, WEB OF SCIENCE and COCHRANE databases from the drugs’ inception to May 2016 to identify clinical trials. Severe adverse events (AEs) (grade ≥ 3) based on the ALK-TKI type were analysed. Seventeen trials published between 2011 and 2016, including a total of 1826 patients, were eligible for analysis. Patients in 10 trials (n = 1000) received crizotinib, patients in 5 trials (n = 601) received ceritinib and patients in 2 trials (n = 225) received alectinib. The overall frequencies of treatment-related death and AEs due to treatment withdrawal were 0.9% (12/1365) and 5.5% (85/1543), respectively. Moreover, the frequency of severe AEs in patients treated with ceritinib was significantly higher than patients treated with crizotinib or alectinib, especially for hepatotoxicity, fatigue and some of gastrointestinal symptoms. Additionally, significant difference in the elevated lipase and amylase levels (grade ≥ 3) were detected between ceritinib and crizotinib/alectinib, whereas neutropenia was less frequent. ALK-TKIs were safe for ALK-positive patients. Moreover, statistically significant differences in some severe AEs among ceritinib, crizotinib and alectinib were detected in present study.
- 12 June 2017
- Published by Springer Nature in BMC Cancer
- Vol. 17 (1) , 1-10
Open AccessPooled safety analyses of ALK-TKI inhibitor in ALK-positive NSCLC
The anaplastic lymphoma kinase tyrosine kinase inhibitors (ALK-TKIs) have been administered to patients with ALK-positive non-small cell lung cancer for a long period of time and show a promising response. However, the differences in the toxicity profiles among these drugs are still unclear. We performed a comprehensive search of the MEDLINE, EMBASE, WEB OF SCIENCE and COCHRANE databases from the drugs’ inception to May 2016 to identify clinical trials. Severe adverse events (AEs) (grade ≥ 3) based on the ALK-TKI type were analysed. Seventeen trials published between 2011 and 2016, including a total of 1826 patients, were eligible for analysis. Patients in 10 trials (n = 1000) received crizotinib, patients in 5 trials (n = 601) received ceritinib and patients in 2 trials (n = 225) received alectinib. The overall frequencies of treatment-related death and AEs due to treatment withdrawal were 0.9% (12/1365) and 5.5% (85/1543), respectively. Moreover, the frequency of severe AEs in patients treated with ceritinib was significantly higher than patients treated with crizotinib or alectinib, especially for hepatotoxicity, fatigue and some of gastrointestinal symptoms. Additionally, significant difference in the elevated lipase and amylase levels (grade ≥ 3) were detected between ceritinib and crizotinib/alectinib, whereas neutropenia was less frequent. ALK-TKIs were safe for ALK-positive patients. Moreover, statistically significant differences in some severe AEs among ceritinib, crizotinib and alectinib were detected in present study.
- Open Access
- 30 September 2016
- Published by China Science Publishing & Media Ltd. in Acta Biochimica et Biophysica Sinica
- Vol. 48 (10) , 948-957
Structure features of GH10 xylanase from Caldicellulosiruptor bescii: implication for its thermophilic adaption and substrate binding preference
Caldicellulosiruptor bescii is the most thermophilic cellulolytic species of organisms known to date. In our previous study, GH10 xylanase CbXyn10B from C. bescii displayed outstanding hydrolytic activity toward various xylans at high temperatures. To understand the structural basis for this protein's catalysis and thermostability, we solved the crystal structures of CbXyn10B and its complexes with xylooligosaccharides. These structural models were used to guide comparison with its mesophilic counterpart PbXyn10B. A distinctive structural feature is that thermophilic CbXyn10B presents a relatively stable interaction between the extended loops L7 and L8 in the catalytic cleft by an extensive hydrogen bonding network, which is mediated by Lys306, Arg314 and three well-ordered water molecules. Moreover, a unique aromatic cluster consisting of Try17, Phe20, Phe21, and Phe337 may enhance the interaction between the N- and C- terminus. Targeted mutagenesis demonstrated that these interactions substantially contribute to enzyme stabilization, as indicated by a considerable decrease in the melting temperature (Tm) of CbXyn10B by substituting critical residues with Ala. Therefore, it was shown that not only the aromatic interaction connecting protein termini but also the extensive hydrogen bonding network formed between surface loops could restrict the local structural flexibility and contribute significantly to the overall stability of enzymes. Furthermore, the xylooligosaccharides were found to tightly bind to the glycone subsites of xylanase, indicating higher affinities at these subsites and reflecting its substrate binding preference. Our results suggest that CbXyn10B is stabilized with distinct rigidity at the catalytic cleft as well as the terminal regions, which provides insights into the evolutionary strategy for accommodating the functional needs of GH10 enzymes to high temperature.
- 30 September 2016
- Published by China Science Publishing & Media Ltd. in Acta Biochimica et Biophysica Sinica
- Vol. 48 (10) , 948-957
Open AccessStructure features of GH10 xylanase from Caldicellulosiruptor bescii: implication for its thermophilic adaption and substrate binding preference
Caldicellulosiruptor bescii is the most thermophilic cellulolytic species of organisms known to date. In our previous study, GH10 xylanase CbXyn10B from C. bescii displayed outstanding hydrolytic activity toward various xylans at high temperatures. To understand the structural basis for this protein's catalysis and thermostability, we solved the crystal structures of CbXyn10B and its complexes with xylooligosaccharides. These structural models were used to guide comparison with its mesophilic counterpart PbXyn10B. A distinctive structural feature is that thermophilic CbXyn10B presents a relatively stable interaction between the extended loops L7 and L8 in the catalytic cleft by an extensive hydrogen bonding network, which is mediated by Lys306, Arg314 and three well-ordered water molecules. Moreover, a unique aromatic cluster consisting of Try17, Phe20, Phe21, and Phe337 may enhance the interaction between the N- and C- terminus. Targeted mutagenesis demonstrated that these interactions substantially contribute to enzyme stabilization, as indicated by a considerable decrease in the melting temperature (Tm) of CbXyn10B by substituting critical residues with Ala. Therefore, it was shown that not only the aromatic interaction connecting protein termini but also the extensive hydrogen bonding network formed between surface loops could restrict the local structural flexibility and contribute significantly to the overall stability of enzymes. Furthermore, the xylooligosaccharides were found to tightly bind to the glycone subsites of xylanase, indicating higher affinities at these subsites and reflecting its substrate binding preference. Our results suggest that CbXyn10B is stabilized with distinct rigidity at the catalytic cleft as well as the terminal regions, which provides insights into the evolutionary strategy for accommodating the functional needs of GH10 enzymes to high temperature.
- 14 September 2016
- Published by Springer Nature in Multidimensional Systems and Signal Processing
- Vol. 29 (1) , 35-47
Sparse representation based two-dimensional direction of arrival estimation using co-prime array
- 14 September 2016
- Published by Springer Nature in Multidimensional Systems and Signal Processing
- Vol. 29 (1) , 35-47
Sparse representation based two-dimensional direction of arrival estimation using co-prime array
- Open Access
- 23 February 2016
- Published by Impact Journals, LLC in Oncotarget
- Vol. 7 (13) , 16248-16261
Bromodomain-containing protein 7 (BRD7) as a potential tumor suppressor in hepatocellular carcinoma
// <![CDATA[ $('.header-date').hide();$('#titleAuthors').hide();$('#abstractHeader').hide(); // ]]> Chang-Long Chen1,2,*, Ying Wang1,3,*, Qiu-Zhong Pan1,2, Yan Tang1,2, Qi-Jing Wang2, Ke Pan1,2, Li-Xi Huang2, Jia He2, Jing-Jing Zhao1,2, Shan-Shan Jiang1,2, Xiao-Fei Zhang1,2, Hong-Xia Zhang1,2, Zi-Qi Zhou1,2, De-Sheng Weng1,2 and Jian-Chuan Xia1,2 1 Collaborative Innovation Center for Cancer Medicine, State Key Laboratory of Oncology in South China, Sun Yat-Sen University Cancer Center, Guangzhou, China 2 Department of Biotherapy, Sun Yat-Sen University Cancer Center, Guangzhou, China 3 Department of Epidemiology and Health Statistics, Guangdong Key Laboratory of Molecular Epidemiology, Guangdong Pharmaceutical University, Guangzhou, China * These authors have contributed equally to this work Correspondence to: Jian-Chuan Xia, email: // <![CDATA[ mail("xiajch","mail.sysu.edu",23,"") // ]]> De-Sheng Weng, email: // <![CDATA[ mail("wengdsh","sysucc.org",23,"") // ]]> Keywords: BRD7, tumor suppressor, hepatocellular carcinoma, prognosis, tumorigenicity Received: December 15, 2015 Accepted: February 08, 2016 Published: February 23, 2016
- 23 February 2016
- Published by Impact Journals, LLC in Oncotarget
- Vol. 7 (13) , 16248-16261
Open AccessBromodomain-containing protein 7 (BRD7) as a potential tumor suppressor in hepatocellular carcinoma
// <![CDATA[ $('.header-date').hide();$('#titleAuthors').hide();$('#abstractHeader').hide(); // ]]> Chang-Long Chen1,2,*, Ying Wang1,3,*, Qiu-Zhong Pan1,2, Yan Tang1,2, Qi-Jing Wang2, Ke Pan1,2, Li-Xi Huang2, Jia He2, Jing-Jing Zhao1,2, Shan-Shan Jiang1,2, Xiao-Fei Zhang1,2, Hong-Xia Zhang1,2, Zi-Qi Zhou1,2, De-Sheng Weng1,2 and Jian-Chuan Xia1,2 1 Collaborative Innovation Center for Cancer Medicine, State Key Laboratory of Oncology in South China, Sun Yat-Sen University Cancer Center, Guangzhou, China 2 Department of Biotherapy, Sun Yat-Sen University Cancer Center, Guangzhou, China 3 Department of Epidemiology and Health Statistics, Guangdong Key Laboratory of Molecular Epidemiology, Guangdong Pharmaceutical University, Guangzhou, China * These authors have contributed equally to this work Correspondence to: Jian-Chuan Xia, email: // <![CDATA[ mail("xiajch","mail.sysu.edu",23,"") // ]]> De-Sheng Weng, email: // <![CDATA[ mail("wengdsh","sysucc.org",23,"") // ]]> Keywords: BRD7, tumor suppressor, hepatocellular carcinoma, prognosis, tumorigenicity Received: December 15, 2015 Accepted: February 08, 2016 Published: February 23, 2016