Abstract
Most cameras today capture images without considering scene content. In contrast, animal eyes have fast mechanical movements that control how the scene is imaged in detail by the fovea, where visual acuity is highest. This concentrates computational (i.e. neuronal) resources in places where they are most needed. The prevalence of foveation, and the wide variety of it, makes it very clear that this is an effective visual design strategy. If robotic drones had this kind of low power vision, we could imagine massive impact on a variety of fields. Allowing these small devices to reliably sense their surroundings has the potential for a major transformation in computer vision. In this talk, I cover our recent work on creating cameras and algorithms that enable fast, selective imaging. The key challenge is that these algorithms require novel camera designs with fast and low-power control of the physical properties that influence image formation. I will discuss examples of such cameras, LIDARs and projectors that utilize newly available, next generation micro-mechanical optics.
© 2020 The Author(s)
PDF Article | Presentation VideoMore Like This
Akshat Dave, Muralidhar Madabhushi Balaji, Prasanna Rangarajan, Ashok Veeraraghavan, and Marc P Christensen
CTh5C.6 Computational Optical Sensing and Imaging (COSI) 2020
Roelof Jansen, Mathias Prost, Jon Kjellman, Aleksandrs Marinins, Philippe Soussan, Marcus S. Dahlem, and Xavier Rottenberg
ITh2H.5 Integrated Photonics Research, Silicon and Nanophotonics (IPR) 2020
Ryan Chu, Dene Ringuette, Eric Zhu, Alex Mihailidis, and Ofer Levi
IW3D.3 Imaging Systems and Applications (IS) 2020