Oral presentation in CVPR 2019. Better stereo view extrapolation by considering sampling and 3D inpainting occluded stuff.

CVPR 2018. Defocus blur provides a supervisory signal for training deep networks to estimate depth from a single all-in-focus image.

SIGGRAPH Asia 2017. Considering the eye’s aberrations when rendering increases realism and helps drive accomodation in VR.

Spotlight Presentation in ICCV 2017. Deep networks can promote a single image to a local neighborhood of views and depth maps.

Oral Presentation in CVPR 2017. Fourier theory and an algorithm to deblur light fields captured with unknown camera motion.

ICCV 2015. 4D light field descriptors to compute better dense scene flow.



Student Researcher

Google AI/Research

Jun 2018 – Present New York, NY
Computer vision and graphics research. Hosted by Professor Noah Snavely. Work published at CVPR 2019.

Student Researcher

Google AI/Research

May 2017 – Nov 2017 Mountain View, CA
Computer vision and computational photography research in Marc Levoy’s team, hosted by Jon Barron. Work published at CVPR 2018.

Graduate Researcher

UC Berkeley

Sep 2014 – Present Berkeley, CA
Research at the intersection of computer vision, computer graphics, and deep learning. Advised by Professor Ren Ng and Professor Ravi Ramamoorthi.

Undergraduate Researcher

Duke University

Jan 2011 – May 2014 Durham, NC
Medical computer vision research. Advised by Professor Sina Farsiu. Received Pratt Research Fellowship and Howard G. Clark Award.