Doctoral Thesis Oral Defense - Dorian Yao Chan

— 12:30pm

Location:
In Person and Virtual - ET - Gates Hillman 8102 and Zoom

Speaker:
DORIAN YAO CHAN , Ph.D. Candidate, Computer Science Department, Carnegie Mellon University
https://dorianchan.com/

Holographic Displays for Computer Vision

Artificial illumination is ubiquitous in real vision systems. By coding extra information from a controlled light source into the images captured by a camera, so-called "active sensing" approaches robustly capture depth, reflectance and other visual cues crucial to tasks in robotics, manufacturing, consumer products and more. However, active sensors struggle with well-known challenges that limit their practicality in modern applications — available power limits range and outdoor performance, slow speed precludes dynamic scenes and defocused structured illumination reduces effective resolution. 

To tackle these challenges, this thesis explores using holographic displays. Holographic displays have recently seen significant attention in the augmented and virtual reality (AR/VR) literature. By simply illuminating a spatial-light modulator (SLM) with laser light, such devices can simultaneously provide accommodation cues and glasses-free vision correction all in a compact form factor, key capabilities that are currently missing in modern AR/VR architectures.

In our work, we analyze how they can potentially be adapted as sources of active illumination. First, we show how holographic displays can be used to build light redistributive projectors that allow for smarter energy usage in active sensing, enabling time-of-flight sensors with far-improved dynamic range. Next, we demonstrate how this light redistribution, when combined with the underlying fast speed of modern SLMs, allows for far faster projector systems, allowing for new types of triangulation light curtains. Finally, we test how the inherent coherent propagation of holographic systems can be used to program meaningful content at multiple projector depths, enabling new user interfaces and depth-sensing methodologies.

Overall, this defense advances the state-of-the-art in active sensing by demonstrating new ways in which light can be shaped and concentrated via holographic illumination systems. These abilities unlock vision systems with increased robustness and newfound capabilities. 

Thesis Committee

Matthew O’Toole (Chair)
Ioannis Gkioulekas
Aswin Sankaranarayanan
Mohit Gupta (University of Wisconsin–Madison)

In Person and Zoom Participation.  See announcement.


Add event to Google
Add event to iCal