Wenbo Wan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Wenbo Wan has a total of 48 co-authors. Here are the top co-authors by number of shared publications:
Qiegen Liu
7 shared publications- Qiegen Liu
Nanchang University
7 shared publications Dingxiang Yuan
3 shared publicationsJiayuan Lin
School of Information Engineering, Nanchang University, Nanchang, China
3 shared publicationsQi Yu
3 shared publicationsTianshui Yu
3 shared publications
Latest Publications
- 20 November 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 206
Measurement domain purification: enabling FZA lensless imaging under complex ambient light interference
- 20 November 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 206
Measurement domain purification: enabling FZA lensless imaging under complex ambient light interference
- 15 April 2026
- Published by Optica Publishing Group in Photonics Research
- Vol. 14 (5) , 1749
High-quality FZA lensless imaging via joint generation of original-twin images
Lensless imaging possesses the advantages of functional flexibility and high portability. The inherent twin-image problem in Fresnel zone aperture (FZA) lensless imaging cannot be eliminated using conventional image reconstruction methods based on spatial sparsity priors. A low-cost FZA encoder with orthogonal phases is used to construct a lensless imaging system, and high-quality FZA lensless imaging via joint generation of original-twin images (LOTI) is proposed. Image generation is performed in a high-dimensional solution space of original-twin images, separating the twin image from the reconstructed content and producing an artifact-free original image. LOTI demonstrates strong robustness to different mask fabrication methods, phase encoding schemes, target distances, and distance mismatch. Simulative and experimental validation demonstrate LOTI effectively suppresses twin-image artifacts in the reconstructed images while preserving textural details and color fidelity. Quantitative evaluations further confirm the superior performance, with an average peak signal-to-noise ratio (PSNR) of 24.37 dB and a structural similarity index measure (SSIM) of 0.92 within the regions of interest.
- 15 April 2026
- Published by Optica Publishing Group in Photonics Research
- Vol. 14 (5) , 1749
High-quality FZA lensless imaging via joint generation of original-twin images
Lensless imaging possesses the advantages of functional flexibility and high portability. The inherent twin-image problem in Fresnel zone aperture (FZA) lensless imaging cannot be eliminated using conventional image reconstruction methods based on spatial sparsity priors. A low-cost FZA encoder with orthogonal phases is used to construct a lensless imaging system, and high-quality FZA lensless imaging via joint generation of original-twin images (LOTI) is proposed. Image generation is performed in a high-dimensional solution space of original-twin images, separating the twin image from the reconstructed content and producing an artifact-free original image. LOTI demonstrates strong robustness to different mask fabrication methods, phase encoding schemes, target distances, and distance mismatch. Simulative and experimental validation demonstrate LOTI effectively suppresses twin-image artifacts in the reconstructed images while preserving textural details and color fidelity. Quantitative evaluations further confirm the superior performance, with an average peak signal-to-noise ratio (PSNR) of 24.37 dB and a structural similarity index measure (SSIM) of 0.92 within the regions of interest.
- 31 March 2026
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 202
Coding-domain prior-aided FZA lensless imaging under finite-sampling
- 31 March 2026
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 202
Coding-domain prior-aided FZA lensless imaging under finite-sampling
- Open Access
- 21 March 2026
- Published by Elsevier in SSRN Electronic Journal
Intensity-guided dual-domain generative reconstruction for coherent diffraction imaging
Coherent diffraction imaging is a representative lensless technique for nanoscale structural characterization by reconstructing complex wavefields from far-field diffraction intensities. However, phase retrieval from intensity-only measurements is inherently nonlinear and severely ill-posed, frequently leading to artifacts and unstable localization when imaging structurally complex targets. To address these challenges, an intensity-guided dual-domain generative reconstruction for coherent diffraction imaging (IDGC) framework is proposed. In the frequency domain, the diffraction intensity measurement guides a mean-reverting diffusion model to directly generate the complex Fourier-domain field. Subsequently, a spatial-domain mean-reverting diffusion model is utilized to suppress residual noise and artifacts, thereby enhancing image fidelity. Simulative and experimental validations demonstrate that IDGC yields reconstructions superior to representative iterative algorithms and existing diffusion-based approaches, characterized by preserved structural details and minimized artifacts. On the LSUN dataset, IDGC achieves a mean peak signal-to-noise ratio of 38.99 dB and a structural similarity index of 0.901, confirming its capability for high-quality reconstruction of structurally complex objects with continuous grayscale variations.
- 21 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessIntensity-guided dual-domain generative reconstruction for coherent diffraction imaging
Coherent diffraction imaging is a representative lensless technique for nanoscale structural characterization by reconstructing complex wavefields from far-field diffraction intensities. However, phase retrieval from intensity-only measurements is inherently nonlinear and severely ill-posed, frequently leading to artifacts and unstable localization when imaging structurally complex targets. To address these challenges, an intensity-guided dual-domain generative reconstruction for coherent diffraction imaging (IDGC) framework is proposed. In the frequency domain, the diffraction intensity measurement guides a mean-reverting diffusion model to directly generate the complex Fourier-domain field. Subsequently, a spatial-domain mean-reverting diffusion model is utilized to suppress residual noise and artifacts, thereby enhancing image fidelity. Simulative and experimental validations demonstrate that IDGC yields reconstructions superior to representative iterative algorithms and existing diffusion-based approaches, characterized by preserved structural details and minimized artifacts. On the LSUN dataset, IDGC achieves a mean peak signal-to-noise ratio of 38.99 dB and a structural similarity index of 0.901, confirming its capability for high-quality reconstruction of structurally complex objects with continuous grayscale variations.
- 14 January 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 196
Undersampled FZA lensless imaging via dual diffusion models with intermediate prior learning
- 14 January 2026
- Published by Elsevier in Optics & Laser Technology
- Vol. 196
Undersampled FZA lensless imaging via dual diffusion models with intermediate prior learning
- 26 October 2025
- Published by Elsevier in Optics Communications
- Vol. 597
TCWD: Temporal compressive coherent diffraction imaging enhanced by weighted wavelet domain diffusion model
- 26 October 2025
- Published by Elsevier in Optics Communications
- Vol. 597
TCWD: Temporal compressive coherent diffraction imaging enhanced by weighted wavelet domain diffusion model
- 15 July 2025
- Published by Elsevier in Optics & Laser Technology
- Vol. 192
Wavelet-refinement-inspired diffusion model for scattering imaging
- 15 July 2025
- Published by Elsevier in Optics & Laser Technology
- Vol. 192
Wavelet-refinement-inspired diffusion model for scattering imaging
- Open Access
- 1 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Photonics Journal
- Vol. 17 (4) , 1-8
Mean-Reverting Diffusion Model-Enhanced Scattering Imaging
Scattering media disrupt the rectilinear propagation of light, significantly degrading the resolution and clarity of optical imaging systems. Current scattering imaging techniques usually focus on simple targets and present limitations in imaging quality and reconstruction efficiency. To address these limitations, a mean-reverting diffusion model-enhanced scattering imaging (MRDS) is proposed. During training, prior information is extracted by diffusing the training data into an intermediate state with stable Gaussian noise. Reconstruction begins with low-quality images from physically-guided inversion, followed by iterative solving of reverse-time stochastic differential equations via the Euler-Maruyama method, integrating learned prior information to efficiently reconstruct high-quality images. Simulative and experimental validations demonstrate that MRDS outperforms traditional methods in reconstructing images with fewer artifacts and enhanced detail clarity. Quantitative metrics further demonstrates excellent reconstruction performance, with average metrics reaching 41.19 dB for PSNR, 0.99 for SSIM and 0.0085 for LPIPS. The reconstruction time per image is 2.19 seconds, representing a 44.2-fold acceleration compared to conventional methods. The proposed method achieves high-quality reconstructions of complex targets in a significantly shorter time, which dramatically boosts the efficiency of scattering imaging.
- 1 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Photonics Journal
- Vol. 17 (4) , 1-8
Open AccessMean-Reverting Diffusion Model-Enhanced Scattering Imaging
Scattering media disrupt the rectilinear propagation of light, significantly degrading the resolution and clarity of optical imaging systems. Current scattering imaging techniques usually focus on simple targets and present limitations in imaging quality and reconstruction efficiency. To address these limitations, a mean-reverting diffusion model-enhanced scattering imaging (MRDS) is proposed. During training, prior information is extracted by diffusing the training data into an intermediate state with stable Gaussian noise. Reconstruction begins with low-quality images from physically-guided inversion, followed by iterative solving of reverse-time stochastic differential equations via the Euler-Maruyama method, integrating learned prior information to efficiently reconstruct high-quality images. Simulative and experimental validations demonstrate that MRDS outperforms traditional methods in reconstructing images with fewer artifacts and enhanced detail clarity. Quantitative metrics further demonstrates excellent reconstruction performance, with average metrics reaching 41.19 dB for PSNR, 0.99 for SSIM and 0.0085 for LPIPS. The reconstruction time per image is 2.19 seconds, representing a 44.2-fold acceleration compared to conventional methods. The proposed method achieves high-quality reconstructions of complex targets in a significantly shorter time, which dramatically boosts the efficiency of scattering imaging.
- 4 February 2025
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 187
Fluorescence molecular tomography via score-based generative model
- 4 February 2025
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 187
Fluorescence molecular tomography via score-based generative model
- 19 August 2024
- Published by Optica Publishing Group in Optics Express
- Vol. 32 (18) , 31920-31938
DMEDH: diffusion model-boosted multiplane extrapolation for digital holographic reconstruction
Digital holography can reconstruct the amplitude and phase information of the target light field. However, the reconstruction quality is largely limited by the size of the hologram. Multi-plane holograms can impose constraints for reconstruction, yet the quality of the reconstructed images continues to be restricted owing to the deficiency of effective prior information constraints. To attain high-quality image reconstruction, a diffusion model-boosted multiplane extrapolation for digital holographic reconstruction (DMEDH) algorithm is proposed. The dual-channel prior information of amplitude and phase extracted through denoising score matching is employed to constrain the physically driven dual-domain rotational iterative process. Depending on the utilization of multi-plane hologram data, the serial DMEDH and the parallel DMEDH are presented. Compared with traditional methods, simulative and experimental results demonstrate that images reconstructed using DMEDH exhibit better reconstruction quality and have higher structural similarity, peak signal-to-noise ratios, and strong generalization. The reconstructed image using DMEDH from two holograms exhibits better quality than that of traditional methods from five holograms.
- 19 August 2024
- Published by Optica Publishing Group in Optics Express
- Vol. 32 (18) , 31920-31938
DMEDH: diffusion model-boosted multiplane extrapolation for digital holographic reconstruction
Digital holography can reconstruct the amplitude and phase information of the target light field. However, the reconstruction quality is largely limited by the size of the hologram. Multi-plane holograms can impose constraints for reconstruction, yet the quality of the reconstructed images continues to be restricted owing to the deficiency of effective prior information constraints. To attain high-quality image reconstruction, a diffusion model-boosted multiplane extrapolation for digital holographic reconstruction (DMEDH) algorithm is proposed. The dual-channel prior information of amplitude and phase extracted through denoising score matching is employed to constrain the physically driven dual-domain rotational iterative process. Depending on the utilization of multi-plane hologram data, the serial DMEDH and the parallel DMEDH are presented. Compared with traditional methods, simulative and experimental results demonstrate that images reconstructed using DMEDH exhibit better reconstruction quality and have higher structural similarity, peak signal-to-noise ratios, and strong generalization. The reconstructed image using DMEDH from two holograms exhibits better quality than that of traditional methods from five holograms.