Hao Sha
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
Hao Sha has a total of 87 co-authors. Here are the top co-authors by number of shared publications:
- Yongbing Zhang
Harbin Institute of Technology (Shenzhen)
6 shared publications - Xiaochen Feng
School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen), Shenzhen, China
5 shared publications - Shangguo Hou
Institute of Systems and Physical Biology, Shenzhen Bay Laboratory, Shenzhen, China
3 shared publications Taiqin Chen
School of Computer Science and Technology Harbin Institute of Technology (Shenzhen) Shenzhen China
3 shared publicationsYuan Jiang
Harbin Institute of Technology
3 shared publications- Chunyu Chen
Faculty of Computing Harbin Institute of Technology Harbin China
2 shared publications
Latest Publications
- Open Access
- 28 October 2026
- Published by Wiley in Advanced Science
ReLink: Reliable Deep Learning for Microscopic Motion Tracking
Single‐particle tracking (SPT) serves as a fundamental technique for deciphering complex microscopic dynamics in biological and physical systems by resolving trajectories of individual particles. However, linking localizations into trajectories remains challenging in realistic experiments, where crowding, blinking, and localization errors lead to linking ambiguity. Both statistical and deep‐learning approaches have been explored in these conditions, yet reliable link selection that generalizes across varying data distributions remains elusive. Here, we propose ReLink, a general deep‐learning framework for reliable trajectory linking in SPT. By integrating physical priors into geometric representations and leveraging confidence‐based adaptive filtering at inference, ReLink enables automated selection of trustworthy linkages, thereby preserving the fidelity of downstream trajectory‐based analyses. Across challenging simulated scenarios spanning broad experimental conditions and motion regimes, ReLink consistently outperforms representative statistical and deep learning‐based trajectory linking methods in both edge‐level and trajectory‐level evaluation metrics, leading to markedly improved downstream inference on public benchmarks. We further validate ReLink on diverse real‐world experiments, from free diffusion systems to crowded intracellular environments, demonstrating its practical significance for quantitative interpretation of microscopic dynamics and its promise for next‐generation trustworthy and automated SPT.
- 28 October 2026
- Published by Wiley in Advanced Science
Open AccessReLink: Reliable Deep Learning for Microscopic Motion Tracking
Single‐particle tracking (SPT) serves as a fundamental technique for deciphering complex microscopic dynamics in biological and physical systems by resolving trajectories of individual particles. However, linking localizations into trajectories remains challenging in realistic experiments, where crowding, blinking, and localization errors lead to linking ambiguity. Both statistical and deep‐learning approaches have been explored in these conditions, yet reliable link selection that generalizes across varying data distributions remains elusive. Here, we propose ReLink, a general deep‐learning framework for reliable trajectory linking in SPT. By integrating physical priors into geometric representations and leveraging confidence‐based adaptive filtering at inference, ReLink enables automated selection of trustworthy linkages, thereby preserving the fidelity of downstream trajectory‐based analyses. Across challenging simulated scenarios spanning broad experimental conditions and motion regimes, ReLink consistently outperforms representative statistical and deep learning‐based trajectory linking methods in both edge‐level and trajectory‐level evaluation metrics, leading to markedly improved downstream inference on public benchmarks. We further validate ReLink on diverse real‐world experiments, from free diffusion systems to crowded intracellular environments, demonstrating its practical significance for quantitative interpretation of microscopic dynamics and its promise for next‐generation trustworthy and automated SPT.
- Open Access
- 22 October 2026
- Published by Wiley in Advanced Science
Snapshot Spatiotemporal‐Spectral Imaging Enabled by Joint Photonic Encoding
Hyperspectral imaging (HSI) enables simultaneous acquisition of spectral and spatial information and has been widely adopted across science and engineering. However, existing HSI systems are primarily designed for x‐y‐λ datacube acquisition, with the temporal dimension insufficiently incorporated and limiting their ability to capture rapid variations in spatial structures or spectral signatures. Here we report Spectra4D (x, y, λ, t), a snapshot spatiotemporal‐spectral imaging architecture that enables high‐dimensional and high‐resolution observation of dynamic scenes beyond the native camera frame‐rate limit. Enabled by joint photonic encoding approach, Spectra4D synergizes high‐speed temporal modulation with on‐sensor spectral multiplexing to jointly encode spatiotemporal and spectral information into a single compressed exposure, which is computationally reconstructed using a dual‐mask guided reconstruction network. Experiments demonstrate that Spectra4D reconstructs 28 visible spectral channels under 4× temporal compression, achieving spectral fidelity above 96.5% and pixel‐level localization accuracy across continuous multi‐frame sequences. We further validate the capability of Spectra4D in dynamic spatio‐spectral scenarios, including dual‐color fluorescent microsphere tracking and frame‐resolved elemental monitoring in flame combustion. In summary, Spectra4D presents a practical and scalable route toward temporally resolved spatio‐spectral imaging of dynamic processes with rapidly varying spectral features.
- 22 October 2026
- Published by Wiley in Advanced Science
Open AccessSnapshot Spatiotemporal‐Spectral Imaging Enabled by Joint Photonic Encoding
Hyperspectral imaging (HSI) enables simultaneous acquisition of spectral and spatial information and has been widely adopted across science and engineering. However, existing HSI systems are primarily designed for x‐y‐λ datacube acquisition, with the temporal dimension insufficiently incorporated and limiting their ability to capture rapid variations in spatial structures or spectral signatures. Here we report Spectra4D (x, y, λ, t), a snapshot spatiotemporal‐spectral imaging architecture that enables high‐dimensional and high‐resolution observation of dynamic scenes beyond the native camera frame‐rate limit. Enabled by joint photonic encoding approach, Spectra4D synergizes high‐speed temporal modulation with on‐sensor spectral multiplexing to jointly encode spatiotemporal and spectral information into a single compressed exposure, which is computationally reconstructed using a dual‐mask guided reconstruction network. Experiments demonstrate that Spectra4D reconstructs 28 visible spectral channels under 4× temporal compression, achieving spectral fidelity above 96.5% and pixel‐level localization accuracy across continuous multi‐frame sequences. We further validate the capability of Spectra4D in dynamic spatio‐spectral scenarios, including dual‐color fluorescent microsphere tracking and frame‐resolved elemental monitoring in flame combustion. In summary, Spectra4D presents a practical and scalable route toward temporally resolved spatio‐spectral imaging of dynamic processes with rapidly varying spectral features.
- 7 October 2026
- Published by American Chemical Society (ACS) in ACS Photonics
- Vol. 13 (20) , 6114-6123
Spectrally Filtered Optical Convolutional Neural Network for Shift-Robust Bright-Field Classification
Optical neural networks (ONNs) offer high-throughput and energy-efficient computation, providing a promising alternative to conventional electronic systems. However, their application in bright-field microscopy is limited by random spatial displacements of dynamic targets and a strong unmodulated transmitted-light component concentrated at the zero order of the spatial-frequency domain. This direct current (DC) component significantly reduces feature contrast, posing a major challenge for reliable optical inference. To address these challenges, we propose a spectrally filtered optical convolutional neural network (SF-OCNN) that integrates Fourier filtering, optical convolution, and an expanded detection region. The Fourier filter suppresses the DC component before optical convolution while preserving high-frequency structural information. Optical convolution provides inherent translation equivariance, while the expanded detection region accommodates the displacement of the output response during inference. The proposed system is validated through both numerical simulations and physical experiments. It achieves validation accuracies of 91.12% in simulation and 88.50% in the physical experiment on a bright-field MNIST data set (digits 0–3). For experimentally acquired microfluidic cell images re-encoded on the DMD, the corresponding accuracies are 99.50% and 98.25%. These results demonstrate that combining physical DC suppression with optical convolution provides an effective approach for robust optical computing in realistic bright-field environments.
- 7 October 2026
- Published by American Chemical Society (ACS) in ACS Photonics
- Vol. 13 (20) , 6114-6123
Spectrally Filtered Optical Convolutional Neural Network for Shift-Robust Bright-Field Classification
Optical neural networks (ONNs) offer high-throughput and energy-efficient computation, providing a promising alternative to conventional electronic systems. However, their application in bright-field microscopy is limited by random spatial displacements of dynamic targets and a strong unmodulated transmitted-light component concentrated at the zero order of the spatial-frequency domain. This direct current (DC) component significantly reduces feature contrast, posing a major challenge for reliable optical inference. To address these challenges, we propose a spectrally filtered optical convolutional neural network (SF-OCNN) that integrates Fourier filtering, optical convolution, and an expanded detection region. The Fourier filter suppresses the DC component before optical convolution while preserving high-frequency structural information. Optical convolution provides inherent translation equivariance, while the expanded detection region accommodates the displacement of the output response during inference. The proposed system is validated through both numerical simulations and physical experiments. It achieves validation accuracies of 91.12% in simulation and 88.50% in the physical experiment on a bright-field MNIST data set (digits 0–3). For experimentally acquired microfluidic cell images re-encoded on the DMD, the corresponding accuracies are 99.50% and 98.25%. These results demonstrate that combining physical DC suppression with optical convolution provides an effective approach for robust optical computing in realistic bright-field environments.
- Open Access
- 30 September 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-11
Single molecule spectrum dynamics imaging with 3D target-locking tracking
Fluorescence spectra offer rich physicochemical insights into molecular environments and interactions. However, imaging the dynamic fluorescence spectrum of rapidly moving biomolecules, along with their positional dynamics, remains a significant challenge. Here, we report a three-dimensional target-locking-based single-molecule fluorescence Spectrum Dynamics Imaging Microscopy (3D-SpecDIM), a method capable of simultaneously capturing both rapid 3D positional dynamics and physicochemical parameter changing dynamics of the biomolecules with enhanced spectral accuracy, high spectral acquisition speed, single-molecule sensitivity, and high 3D spatiotemporal localization precision. As a demonstration, 3D-SpecDIM is applied to real-time spectral imaging of the mitophagy process, highlighting its enhanced ratiometric fluorescence imaging capability. Additionally, 3D-SpecDIM is used for multi-resolution imaging, providing valuable contextual information on the mitophagy process. Furthermore, we demonstrated the quantitative imaging capability of 3D-SpecDIM by imaging the cellular blebbing process. By continuously monitoring the physicochemical parameter dynamics of biomolecular environments through spectral information, coupled with 3D positional dynamics imaging, 3D-SpecDIM offers a versatile platform for concurrently acquiring multiparameter dynamics, providing comprehensive insights unattainable through conventional imaging techniques. This work represents a substantial advancement in single-molecule spectral dynamics imaging techniques.
- 30 September 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-11
Open AccessSingle molecule spectrum dynamics imaging with 3D target-locking tracking
Fluorescence spectra offer rich physicochemical insights into molecular environments and interactions. However, imaging the dynamic fluorescence spectrum of rapidly moving biomolecules, along with their positional dynamics, remains a significant challenge. Here, we report a three-dimensional target-locking-based single-molecule fluorescence Spectrum Dynamics Imaging Microscopy (3D-SpecDIM), a method capable of simultaneously capturing both rapid 3D positional dynamics and physicochemical parameter changing dynamics of the biomolecules with enhanced spectral accuracy, high spectral acquisition speed, single-molecule sensitivity, and high 3D spatiotemporal localization precision. As a demonstration, 3D-SpecDIM is applied to real-time spectral imaging of the mitophagy process, highlighting its enhanced ratiometric fluorescence imaging capability. Additionally, 3D-SpecDIM is used for multi-resolution imaging, providing valuable contextual information on the mitophagy process. Furthermore, we demonstrated the quantitative imaging capability of 3D-SpecDIM by imaging the cellular blebbing process. By continuously monitoring the physicochemical parameter dynamics of biomolecular environments through spectral information, coupled with 3D positional dynamics imaging, 3D-SpecDIM offers a versatile platform for concurrently acquiring multiparameter dynamics, providing comprehensive insights unattainable through conventional imaging techniques. This work represents a substantial advancement in single-molecule spectral dynamics imaging techniques.
- 27 September 2025
- Published by Elsevier in Information Fusion
- Vol. 127
A channel- adaptive and plug-and- play framework for hyperspectral image analysis
- 27 September 2025
- Published by Elsevier in Information Fusion
- Vol. 127
A channel- adaptive and plug-and- play framework for hyperspectral image analysis
- Top Cited PapersOpen Access
- 22 July 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-17
Quantitative evaluation of methods to analyze motion changes in single-particle experiments
The analysis of live-cell single-molecule imaging experiments can reveal valuable information about the heterogeneity of transport processes and interactions between cell components. These characteristics are seen as motion changes in the particle trajectories. Despite the existence of multiple approaches to carry out this type of analysis, no objective assessment of these methods has been performed so far. Here, we report the results of a competition to characterize and rank the performance of these methods when analyzing the dynamic behavior of single molecules. To run this competition, we implemented a software library that simulates realistic data corresponding to widespread diffusion and interaction models, both in the form of trajectories and videos obtained in typical experimental conditions. The competition constitutes the first assessment of these methods, providing insights into the current limitations of the field, fostering the development of new approaches, and guiding researchers to identify optimal tools for analyzing their experiments.
- 22 July 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-17
Top Cited PapersOpen AccessQuantitative evaluation of methods to analyze motion changes in single-particle experiments
The analysis of live-cell single-molecule imaging experiments can reveal valuable information about the heterogeneity of transport processes and interactions between cell components. These characteristics are seen as motion changes in the particle trajectories. Despite the existence of multiple approaches to carry out this type of analysis, no objective assessment of these methods has been performed so far. Here, we report the results of a competition to characterize and rank the performance of these methods when analyzing the dynamic behavior of single molecules. To run this competition, we implemented a software library that simulates realistic data corresponding to widespread diffusion and interaction models, both in the form of trajectories and videos obtained in typical experimental conditions. The competition constitutes the first assessment of these methods, providing insights into the current limitations of the field, fostering the development of new approaches, and guiding researchers to identify optimal tools for analyzing their experiments.
- Open Access
- 22 May 2025
- Published by IOP Publishing in Journal of Physics: Photonics
- Vol. 7 (3) , 035006
Enhancing diffusion estimation in single-particle experiments through motion change analysis using deep learning
Diffusion is a fundamental process in many scientific disciplines, and accurately characterizing diffusion at the single-molecule level is crucial for understanding complex dynamic systems. To advance and benchmark anomalous diffusion (AnDi) analysis methods in the presence of motion changes, an international team of scientists launched the second AnDi Challenge in March 2024, assessing and comparing new and existing methods for diffusion estimation in dynamic systems. In response to the challenge, we introduce U-LFormer, a deep learning framework designed for precise motion analysis in single-particle tracking (SPT) experiments. By integrating advanced network architecture, training objective, and motion change analysis, U-LFormer excels in change points detection and diffusion parameters recognition across trajectories of various lengths. In the second AnDi Challenge, U-LFormer achieved first place in the Video Track Ensemble task and second place in the trajectory track single-trajectory task, being the only method to consistently rank among the top performers across all tasks. Further evaluations in this study demonstrate the robustness and versatility of U-LFormer, pushing the frontiers of diffusion analysis in SPT experiments.
- 22 May 2025
- Published by IOP Publishing in Journal of Physics: Photonics
- Vol. 7 (3) , 035006
Open AccessEnhancing diffusion estimation in single-particle experiments through motion change analysis using deep learning
Diffusion is a fundamental process in many scientific disciplines, and accurately characterizing diffusion at the single-molecule level is crucial for understanding complex dynamic systems. To advance and benchmark anomalous diffusion (AnDi) analysis methods in the presence of motion changes, an international team of scientists launched the second AnDi Challenge in March 2024, assessing and comparing new and existing methods for diffusion estimation in dynamic systems. In response to the challenge, we introduce U-LFormer, a deep learning framework designed for precise motion analysis in single-particle tracking (SPT) experiments. By integrating advanced network architecture, training objective, and motion change analysis, U-LFormer excels in change points detection and diffusion parameters recognition across trajectories of various lengths. In the second AnDi Challenge, U-LFormer achieved first place in the Video Track Ensemble task and second place in the trajectory track single-trajectory task, being the only method to consistently rank among the top performers across all tasks. Further evaluations in this study demonstrate the robustness and versatility of U-LFormer, pushing the frontiers of diffusion analysis in SPT experiments.
- Open Access
- 1 September 2023
- Published by AIP Publishing in APL Photonics
- Vol. 8 (9)
Deep learning-enhanced single-molecule spectrum imaging
Fluorescence is widely used in biological imaging and biosensing. Rich information can be revealed from the fluorescence spectrum of fluorescent molecules, such as pH, viscosity and polarity of the molecule’s environment, and distance between two FRET molecules. However, constructing the fluorescence spectrum of a single fluorescent molecule typically requires a significant number of photons, which can suffer from photobleaching and, therefore, limit its potential applications. Here, we propose a deep learning-enhanced single-molecule spectrum imaging method (SpecGAN) for improving the single-molecule spectrum imaging efficiency. In SpecGAN, the photon flux required to extract a single-molecule fluorescence spectrum can be reduced by 100 times, which enables two orders of magnitude higher temporal resolution compared to the conventional single-molecule spectrometer. The concept of SpecGAN was validated through numerical simulation and single Nile Red molecule spectrum imaging on support lipid bilayers (SLBs). With SpecGAN, the super-resolution spectrum image of the COS-7 membrane can be reconstructed with merely 12 000 frames of single-molecule localization images, which is almost half of the previously reported frame count for spectrally resolved super-resolution imaging. The low photon flux requirement and high temporal resolution of SpecGAN make it a promising tool for investigating the molecular spectrum dynamics related to biological functions or biomolecule interactions.
- 1 September 2023
- Published by AIP Publishing in APL Photonics
- Vol. 8 (9)
Open AccessDeep learning-enhanced single-molecule spectrum imaging
Fluorescence is widely used in biological imaging and biosensing. Rich information can be revealed from the fluorescence spectrum of fluorescent molecules, such as pH, viscosity and polarity of the molecule’s environment, and distance between two FRET molecules. However, constructing the fluorescence spectrum of a single fluorescent molecule typically requires a significant number of photons, which can suffer from photobleaching and, therefore, limit its potential applications. Here, we propose a deep learning-enhanced single-molecule spectrum imaging method (SpecGAN) for improving the single-molecule spectrum imaging efficiency. In SpecGAN, the photon flux required to extract a single-molecule fluorescence spectrum can be reduced by 100 times, which enables two orders of magnitude higher temporal resolution compared to the conventional single-molecule spectrometer. The concept of SpecGAN was validated through numerical simulation and single Nile Red molecule spectrum imaging on support lipid bilayers (SLBs). With SpecGAN, the super-resolution spectrum image of the COS-7 membrane can be reconstructed with merely 12 000 frames of single-molecule localization images, which is almost half of the previously reported frame count for spectrally resolved super-resolution imaging. The low photon flux requirement and high temporal resolution of SpecGAN make it a promising tool for investigating the molecular spectrum dynamics related to biological functions or biomolecule interactions.
- 12 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 5, 1924-1928
Atmospheric pressure air plasma uniformity study
With the further research of atmospheric pressure low temperature plasma in the field of material treatment, a high frequency high voltage plasma experimental platform is built and a diffuse air plasma at the atmospheric pressure is implemented. A judgement method of plasma uniformity is proposed with the egg inner membrane surface ablative degree after plasma treating. The egg inner membrane. The comparison experiments are set with air and argon plasma as the treating work gas. The result shows that the egg inner membrane surface treating with argon plasma jet causing more serious ablation than the air plasma, because the argon plasma jet obtain discharge filaments while the air plasma does not. Then the electric field distribution in a needle plate electrode system is analyzed with different gap distance through experiments and simulations. and the effect of the gap distance on the discharge uniformity is discussed. According to the electric field distribution results of 5-12mm, it is found that the needle-plate electrode produced plasma that progressed outwards from the needle electrode.
- 12 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 5, 1924-1928
Atmospheric pressure air plasma uniformity study
With the further research of atmospheric pressure low temperature plasma in the field of material treatment, a high frequency high voltage plasma experimental platform is built and a diffuse air plasma at the atmospheric pressure is implemented. A judgement method of plasma uniformity is proposed with the egg inner membrane surface ablative degree after plasma treating. The egg inner membrane. The comparison experiments are set with air and argon plasma as the treating work gas. The result shows that the egg inner membrane surface treating with argon plasma jet causing more serious ablation than the air plasma, because the argon plasma jet obtain discharge filaments while the air plasma does not. Then the electric field distribution in a needle plate electrode system is analyzed with different gap distance through experiments and simulations. and the effect of the gap distance on the discharge uniformity is discussed. According to the electric field distribution results of 5-12mm, it is found that the needle-plate electrode produced plasma that progressed outwards from the needle electrode.