“Light Transport for Computer Vision and Machine Learning”
Tuesday, Dec. 18, 1:00 p.m.
Light transport models the interactions of light paths in a scene with material reflectance and geometry to form images. Traditional computer vision and machine learning algorithms have difficulty with modeling and mitigating light transport effects from 3D scanning to material recognition to non-line-of-sight (NLOS) imaging. In this talk, we will present two different ways in which knowledge of light transport can help improve these applications. First, we will describe a synchronized projector-camera system that captures indirect or multiple bounce light, and use this for improved material recognition and vein visualization. Then we will show how knowledge of light transport can help improve 3D localization of objects outside the line of sight of camera/projector by allowing deep learning algorithms to extract more information from the photons in the NLOS. All this work points to a convergence of physics-based vision with machine learning for robust visual computing systems in the future.
Suren Jayasuriya is an assistant professor at Arizona State University, in the School of Arts, Media and Engineering (AME) and Electrical, Computer and Energy Engineering (ECEE). Before this, he was a postdoctoral fellow at the Robotics Institute at Carnegie Mellon University from 2017-2018. Suren received his Ph.D. in ECE at Cornell University in Jan 2017 and graduated from the University of Pittsburgh in 2012 with a B.S. in Mathematics (with departmental honors) and a B.A. in Philosophy. His research interests are computational imaging and photography, computer vision, and image sensors. He has received the NSF Graduate Research Fellowship in 2013, the Qualcomm Innovation Fellowship in 2015, the ICCV Young Researcher Travel Award in 2017, and the best paper award at the International Conference on Computational Photography in 2014.