I am a research scientist at Facebook Reality Labs where I work on novel display and imaging systems for virtual and augmented reality. I recently graduated with my PhD from the department of Electrical Engineering and Computer Science at UC Berkeley, advised by Professor Laura Waller and Professor Ren Ng.
My research is in computational imaging, which is the joint design of hardware and algorithms for imaging and display systems. I work at the intersection of optics, signal processing, computer graphics, and optimization.
Holographic displays can generate high quality 3D images, but they suffer from a tradeoff between field-of-view and eyebox size. Here, we break the tradeoff for dense, high resolution images using a scattering mask and optimization with perceptually-inspired loss functions.
We present a flat, on-chip, fluorescence microscope, which uses a diffuser in place of a traditional microscope objective. In addition to being inexpensive and compact, our system has a practical working distance, can capture volumetic information in a single shot, and features a new diffuser design that is robust to noise.
Fourier DiffuserScope is a 3D fluorescence microscope capable of capturing large, high-resolution volumes at video rates. We demonstrate that Fourier DiffuserScope outperforms light field microscopy both theoretically and in numerical simulation, and we show results from an experimental prototype.
Lenless cameras can be smaller and less expensive than traditional imaging systems, but the reconstructed images suffer from long compute times and reduced image quality. Here, we improve perfomance by unrolling the traditional model-based optimization algorithm and learning the model parameters from experimentally-captured ground truth data.
We demonstrate an easy-to-build, compact computational camera for single-shot 3D imaging. Our system uses a diffuser (a bumpy piece of plastic) to encode 3D information into a 2D measurement. By using a physical approximation and a simple calibration scheme, we solve for the 3D volume in a computationally efficient way.
Build your own DiffuserCam using a Raspberry Pi camera sensor and scotch tape as the diffuser! Our tutorial includes an overview of DiffuserCam, instructions for putting together your camera, and iPython notebooks for reconstructing images from your raw data.
EECS 16A (Designing Information Devices and and Systems I) is an introductory undergraduate course aimed at incoming electrical engineering and computer science students. The course covers linear algebra, circuits, and signal processing and features hands-on labs. In Summer 2020, we taught the first fully online version of this course, including labs!