Ultrafast Automatic Event Recognition Using Multiphoton Atomic Transitions
Abstract
We will investigate the scientific challenges for realizing an ultrafast automatic event recognition (AER) system using multiphoton atomic transitions, in a manner that is shift, scale and rotation invariant spatially and shift and scale invariant temporally. The goal of the AER is to recognize the presence of a video clip inside a larger video recording. The concept for the AER combines the ideas of spatial image recognition using holography, and temporal signal recognition using stimulated photon echo via four wave mixing. In order to combine these processes, it is necessary to use a non-linear spatio-temporal memory (NOLSTEM) medium that has the requisite type of non-linearity, a high spatial resolution, along with a large inhomogeneous broadening, accompanied by a long coherence lifetime. Since efficient optical transitions are short-lived, it is necessary to use a two-photon transition that can transfer optical information to long-lived spin coherence. Meeting these requirements pose a daunting challenge, requiring fundamental studies, theoretically as well as experimentally. To this end, we will investigate cold 87-Rb atoms held in an array of far off-resonance dipole force traps, in the presence of a magnetic field gradient, as the medium that may potentially meet the requirements of the NOLSTEM. Under this project, we will carry out theoretical and experimental investigations of such a system, for various possible configuration and trap sizes, in order to determine the conditions under which it could be used for realizing the AER system. Specific scientific issues to be investigated include development of optimal cooling and trapping techniques, identification of various sources of dephasing of the spin-coherence and development of techniques for suppression thereof, the effect of radiation trapping in determining the shape of anisotropic atom cloud needed for optimal loading of trap arrays, and effects of quantized motion in traps. Ultrafast Automatic Event Recognition Using Multiphoton Atomic Transitions.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 07, 2024
- Source ID
- FA95502310617
Entities
People
- Mohamd Selim Shahriar
Organizations
- Air Force Office of Scientific Research
- Northwestern University
- United States Air Force