Hello! I am an EECS PhD Candidate at UC Berkeley advised by Laura Waller. I work on computational imaging, which is the joint design of imaging hardware and algorithms, to create more capable cameras and microscopes. My work is at the intesection of signal processing, optics, optimization, and machine learning. I am part of the Berkeley Artificial Intelligence Research (BAIR) Lab.
I completed my B.S. in Electrical Engineering from the State University of New York at Buffalo in May 2016. At Buffalo, I was involved in a nanosatellite mission and several other space-related research projects, which you can read more about on my old website here.
Congrats to my REU mentee Vi Tran for transfering from Orange Coast College to UC Berkeley!
Congrats to my undergraduate mentee Ellin Zhao on choosing UCLA for her PhD!
Congrats to my undergraduate mentee Nico Deshler on choosing the University of Arizona Optics for his PhD!
Selected to participate in the NextProf Nexus 2020 Workshop
Congrats to my undergraduate mentees Ellin and Nico on their paper at OSA's Imaging and Applied Optics Congress on multi-sensor lensless imaging!
Deep learning-based reconstruction methods can improve image quality for many inverse problems, but for high-dimensional imaging (e.g. high speed video, hyperspectral, etc.) obtaining labeled pairs to train deep networks is often impractical or impossible. In this work, we propose to use unsupervised learning for compressive lensless photography. Our 'untrained network' is optimized using only our measurement and physics model to recover a video or hyperspectral volume from a 2D measurement. We demonstrate improved image quality for single-shot compressive video and single-shot hyperspectral imaging without needing any training data.
In this work, we propose a novel, compact, and inexpensive computational camera for snapshot hyperspectral imaging. Our system consists of a repeated spectral filter array placed directly on the image sensor and a diffuser placed close to the sensor. Each point in the world maps to a unique pseudorandom pattern on the spectral filter array, which encodes multiplexed spatio-spectral information. A sparsity-constrained inverse problem solver then recovers the hyperspectral volume with good spatio-spectral resolution. By using a spectral filter array, our hyperspectral imaging framework is flexible and can be designed with contiguous or non-contiguous spectral filters that can be chosen for a given application.
In this work, we replace the tube lens of a Miniscope with an engineered and optimized diffuser that's printed using a Nanoscribe 3D printer. The resulting imager is inexpensive, tiny (the size of a quarter), and can capture 3D fluorescent volumes from a single image, with resulting 3 micron lateral resolution and 10 micron axial resolution at video rates with no moving parts. Check out more of our 3D videos of water bear videos here.
Mask-based lensless imagers, like DiffuserCam, can be small, compact, and capture higher-dimensional information (3D, temporal), but the reconstruction time is slow and the image quality is often degraded. In this work, we show that we can use knowledge of optical system physics along with deep learning to form an unrolled model-based network to solve the reconstruction problem, thereby using physics + deep learning together to speed up and improve image reconstructions. As compared to traditional methods, our architecture achieves better perceptual image quality and runs 20× faster, enabling interactive previewing of the scene.