Speaker
Description
Recent advances in artificial intelligence, compact computing hardware, and
CMOS imaging sensors motivate physics-informed cameras (PhI-Cam). Here different
physics-driven information can come from device-specific sensor materials, sensor
architecture, electronics and noise, environmental parameters including illumination, and
degradation of the hardware performance over time. Given enough data, for example,
from calibration, device modeling, and accumulative use, all such information may be
captured or `learnt’ through a machine learning framework such as a neural-networkbased
deep learning, and implemented through either onboard or off-line hardware
schemes. In this talk, we give an overview about recent advances in high-speed imaging,
experimental needs at the Argonne Advanced Photon Source (APS) synchrotron and its
near future upgrades, and highlight some progress towards PhI-Cam for high-speed Xray
applications. This work is supported in part by LANL C2, C3 and LDRD programs.
