Shuran Song is a second-year PhD student at Princeton University, advised by Professor Jianxiong Xiao. Her research interests are in the areas of computer vision. Specifically, she focuses on data-driven scene understanding in 3D.

Research Summary

Understanding the world in a single glance is one of the most accomplished feats of human intelligence. Yet, despite significant effort made in the past five decades, this scene-understanding problem is still an open challenge for computers. Recently, affordable depth sensors arrived on the consumer market and have made depth acquisition significantly easier. Meanwhile, various crowd-sourcing efforts have resulted in millions of 3D polygonal models available on the Internet.

The combination of these two development completely changes how we can approach the problem. Shuran’s research studies robust ways to leverage these big 3D data to rethink about scene understanding from a novel 3D perspective. The key idea is to go beyond standard 2D pixel-level reasoning and redefine recognition principles in 3D to look for a data-driven scene representation.

For more information, please visit her website: http://vision.princeton.edu/people/shurans/