Research

Our lab develops computational imaging methods, which jointly design optics, devices, signal processing and algorithms, and enable novel capabilities that each one alone cannot. Our research is inherently interdisciplinary, combining expertise in optical engineering, physics and computation. We work on imaging technologies for biomedical, neuroscience, and semiconductor-related applications.


Deep Learning for Computational Imaging 

Implicit neural representation

We develop computational imaging techniques that leverage implicit neural representations.

 

Reliable deep learning with Uncertainty Quantification

We develop uncertainty quantification techniques to provide more reliable deep learning predictions for quantitative bio-imaging.

 

 

Adaptive deep learning

We work on adaptive deep learning framework to achieve robust imaging across a variety of conditions.

 


Computational Microscopy

Computational Miniature Mesoscope (CM2)

We develop “wearable” computational fluorescence microscope that achieves cm-scale FOV and µm-scale resolution with single-shot 3D imaging capability. A complete publication list is here.

 

Event-driven dynamic microscopy

We develop computational “event-driven” microscopy techniques to address challenges in dynamic imaging.

 

 


Computational Semi-conductor Metrology

We developed computational imaging techniques for semiconductor metrology and inspection applications.

Fourier Ptychograhpic Topography

We develop novel topography techniques based on the reflection-mode Fourier ptychographic microscopy, termed Fourier ptychograhpic topography (FPT). FPT provides both a wide FOV and high resolution, and achieves nanoscale height reconstruction accuracy.

 

 

 

 


Computational Phase Imaging

We work in the following major directions. A complete publication list is here.

Intensity Diffraction Tomography

We are developing high-speed computational 3D phase microscopy techniques by leveraging simple optical setups and advanced algorithms.

 

Fourier Ptychography 

We are working computational microscopy techniques to achieve gigapixel phase imaging.

 

 

Differential phase contrast microscopy

We are working computational microscopy techniques based on the principle of transfer function analysis.

 

 

 

Computational label-free chemical microscopy

We are working computational label-free microscopy techniques with chemically specific information.

 

 


Deep learning for Quantitative Bio-imaging

We work in the following major directions.  A complete publication list is here.

Deep learning for in-vivo neural signal modeling & extraction

We develop various deep learning techniques to enhance and analyze in-vivo neural signals. A complete publication list is here.

 

 

Cross-modality information transfer

We work on deep learning technique that allows knowledge transfer across different bio-imaging modalities.

 

 

 


Computational Imaging in Complex Media

We work in the following major directions.  A complete publication list is here.

Multiple-scattering modeling & Physics-based deep learning

We develop efficient and accurate multiple-scattering models to enable large-scale simulation of multiple-scattering in biological samples. These multiple scattering models form the foundation to develop physics-based deep learning models to achieve accurate 3D phase reconstructions. A complete publication list is here.

 

 

 


Computational imaging with non-conventional optics

Computational Imaging with Metasurface Photodetectors

We develop computational imaging techniques for metasurface photodetectors to achieve non-conventional imaging capabilities. A complete publication list is here.