Have a UROP opening you would like to submit?
Please fill out the form.
Enhanced Imaging and Computer Vision enabled by 3D-Printed Nanophotonic Inverse Design
6: Electrical Engineering and Computer Science
Computational imaging and computer vision plays an increasingly important role in modern technology, ranging from simplest image de-noising algorithms to state-of-the-art object recognition, robotic vision and machine intelligence technologies with widespread demand in defense, medical as well as emerging Internet-of-things (IOT) industries. Traditional computer vision is exclusively driven by innovating the computational backend, and more recently, via deep learning and AI software. Little attention has been ever paid to the optical hardware at the front-end beyond conventional lenses and diffraction gratings, which treat light propagation and image formation at a fairly gross level as a primarily geometric problem. The last decade has seen explosive advances in understanding and manipulation of light waves and light-matter interactions at the most profound level of nano-materials, abetted by the development of efficient numerical modeling/design techniques as well as the advent of sophisticated nano-fabrication machinery. In this project, we will seek to facilitate the next-generation computer vision technologies by bringing deeper and richer physics to the optical frontend where the conventional gross hardware will be replaced with exquisitely designed nanophotonic structures, which will not only allow ultra-compact form factors but also enable unprecedented capabilities for physical data acquisition and manipulation. The student will be required to work on very hands-on problems cutting across physics, math and programming as well as commit to some lab time, depending on the skills and propensity of the successful applicant. In particular, the student will work with large-scale photonic inverse design software developed in Prof. Soljacic and Prof. Johnson’s groups, which seamlessly integrates the full vectorial Maxwell equations with computer vision algorithms. More specifically, the student will work on developing an ultra-compact hyperspectral thermal imager capable of simultaneously capturing thermal images and analyzing the radiated spectrum of the captured image in great details. Depending on the time and progress, the student will also be given an opportunity to 3D-print successful designs via nano-scale direct laser printing as well as characterize the fabricated prototypes. A minimum of ~10 hours of work per week is expected from the student on this project, in order to accomplish reasonable progress and learn new skills over the course of the semester.
A solid background in linear algebra, and optics/electromagnetics is required, and some expertise in machine learning or computer vision would be a plus.