2026
Diffusion-Based Joint Recovery, Denoising, and Super-Resolution of Compressed-Sensing Satellite LiDAR Data
IEEE Transactions on Computational Imaging
Research direction
Generative AI for Inverse Problems
Diffusion models, GANs, and uncertainty-aware inference for accelerated, high-fidelity reconstruction in scientific imaging.
Overview
The group brings modern generative models to scientific inverse problems, using diffusion and adversarial priors to recover, denoise, and super-resolve severely undersampled data. Learned priors capture the structure of natural scenes and physical signals, filling in detail that classical regularizers cannot.
Applications include super-resolved 3-D satellite LiDAR, submeter topography and vegetation mapping, and low-dose X-ray CT, together with objective and subjective quality assessment of the generated reconstructions.
Publications
Papers in this direction
2025
Toward Submeter Satellite Surface Topography and Vegetation Mapping Using LiDAR/RGB Constrained Generative Diffusion
IEEE Transactions on Geoscience and Remote Sensing
2025
Super-Resolved 3-D Satellite Lidar Imaging of Earth via Generative Diffusion Models
IEEE Transactions on Geoscience and Remote Sensing
2025
Objective and Subjective Quality Assessment of Forest Landscapes Reconstructed Using Generative Diffusion Models from Compressed Satellite LiDAR Data
IEEE Access
2024
Super-Resolution in Low-Dose X-ray CT via Focal Spot Mitigation with Generative Diffusion Networks
IEEE Transactions on Computational Imaging