Computational Imaging
Developing algorithms and learning-based solutions to improve sensitivity, resolution, and robustness of imaging systems for microscopy and biomedical applications.
Overview#
My work focuses on self-supervised denoising, physically aware reconstruction, and generative priors that enable high-quality imaging under low signal-to-noise conditions. I combine data-driven models with physics constraints to deliver practical improvements for real-world instruments.
Key projects#
- Real-time self-supervised denoising for high-speed fluorescence neural imaging (recovering neural dynamics under low SNR).
- Generative/denoising priors for fluorescence microscopy to reveal subcellular structures and dynamics.
Representative publications#
- “Real‑time Self‑supervised Denoising for High‑speed Fluorescence Neural Imaging” (Nature Communications, 2025)
- “GaMA: Stochastic Gaussian‑masked Denoiser for Enhanced Fluorescence Microscopy of Subcellular Structures and Neural Dynamics” (Advanced Technology in Neuroscience, 2025)
- “Deep Low‑excitation Fluorescence Imaging Enhancement” (Optics Letters, 2022)
If you need code, data, or preprints related to these projects, contact me or check the Publications page for links and resources.