Optical imaging of ultrafast photo-induced phase transitions of nanostructures using neural lenses
Abstract
Ultrafast imaging is crucial to understand the dynamics of light-matter interactions. Recently, an ultrafast iniaging method, termed as Ultrafast Ultrarnicroscopy, was developed in the PI s lab that enabled us to optically in1age ultrafast dynamics of photo-induced surface strnctural evolution. This ultrafast imaging technique spans nine orders of magnitude in the temporal domain from femtosecond to microsecond. However, the spatial resolution is intrinsically diffraction limited. Applying resolution enltancement on ultrafast (non-repeatable) and label-free images is extremely challenging. Recent advancements in neural networks, however, showed the possibility of generating high-resolution images from low-resolution ones. Here, the PI proposes using a Gain Adversarial Neural Nenvork to enltance the resolution of the transient images obtained from our ultrafast ultramicroscope. The proposed neural lens processing will enable high spatial resolution ultrafast dynamic optical in1aging.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 09, 2020
- Source ID
- W911NF2010256
Entities
People
- Chunlei Guo
Organizations
- Army Contracting Command
- United States Army
- University of Rochester