Parsa Mirdehghan

I am a Computer Science PhD candidate at University of Toronto, advised by Kyros Kutulakos, and a Junior Fellow at Massey College. I am also affiliated with Vector Institute, Dynamic Graphics Project (DGP), and Toronto Computational Imaging Group.

My PhD research lies at the intersection of Machine Learning and Computational Imaging, with a focus on developing physics-driven computational models and machine learning algorithms to reveal the underlying attributes of the physical world. By leveraging computational imaging systems that actively encode information into the captured measurements, I explore how to extract 3D geometry, appearance, velocity, and polarization with high accuracy, speed, and robustness. More broadly, I am interested in Computational Imaging and Photography, 3D Computer Vision, Neural Rendering, and Perception.

Previously, I was a research intern at Disney Research Zurich, Huawei Research Canada, and Microsoft Research Redmond.

Twitter  /  Google Scholar /  Linkedin /  Email

We use a coherent optical modem, conventionally used in long-distance data communication networks, and turn it into a new imaging modality, called Full-Wavefield Lidar, for simultaneous measurements of mm-scale depth, velocity, and polarization from only 1 microsecond exposure, at eye-safe regime.

TurboSL provides sub-pixel-accurate surfaces and normals at mega-pixel resolution from structured light images, captured at fractions of a second.

We introduce a method to do novel view lidar synthesis, allowing sparse view scene reconstruction.

We present optical SGD, a computational imaging technique that allows an active depth imaging system to automatically discover optimal illuminations & decoding.

We propose a framework for optimizing structured light patterns given a set of imaging conditions and an error metric.

We present a system specifically designed for capturing the optical properties of live human teeth such that they can be realistically re-rendered in computer graphics

Template from Jon Barron's website.