Shumian Xin

I am a research scientist on Adobe's computational photography team led by Marc Levoy. Previously, I worked with Prof. Ioannis Gkioulekas and Prof. Srinivasa Narasimhan on computational imaging research and received my Ph.D. degree in Robotics from Carnegie Mellon University. I received my Master's degree in Electrical and Computer Engineering from CMU, and my Bachelor's degree in Electrical Engineering from Xi'an Jiaotong University.

Email  /  CV  /  Google Scholar  /  LinkedIn

profile photo
Publications
Defocus Map Estimation and Deblurring from a Single Dual-Pixel Image
Shumian Xin, Neal Wadhwa, Tianfan Xue, Jonathan Barron,
Pratul Srinivasan, Jiawen Chen, Ioannis Gkioulekas, Rahul Garg
IEEE International Conference on Computer Vision (ICCV), 2021
Oral Presentation
Project webpage / Paper / Code

Proposed a single-shot depth from defocus method using commercial dual-pixel sensors, and formulated an optimization problem to jointly recover the defocus map and all-in-focus image of an unknown scene from a single dual-pixel image with potential defocus blurs.

A Theory of Fermat Paths for Non-Line-of-Sight Shape Reconstruction
Shumian Xin, Sotiris Nousias, Kiriakos N. Kutulakos,
Aswin C. Sankaranarayanan, Srinivasa G. Narasimhan, Ioannis Gkioulekas
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019
Oral Presentation, Best Paper Award
Project webpage / Paper / Code / Invited talk

Explored the seemingly impossible NLOS imaging problem of reconstructing objects which are completely occluded and outside the line-of-sight of a camera, proposed a novel theory of Fermat paths for NLOS shape reconstruction, and obtained high-resolution results as if the NLOS objects were directly visible.

Professional Service and Teaching
Journal and Conference Reviewer
  • IEEE Transactions on computational imaging (TCI)
  • Optics Express
  • IEEE International Conference on Computer Vision (ICCV)

  • Student Volunteer
  • IEEE International Conference on Computational Photography (ICCP), 2018 & 2022

  • Teaching Assistant
  • 16-720 Computer Vision, Carnegie Mellon University, Fall 2019 & Spring 2020

  • Template from Jon Barron. Last updated in April 2023.