Image Surveillance Assistant Architecture:Status and Planned Extensions

We are developing a system that integrates methods from
Deep Learning and AI for imagery surveillance. Our
architecture, the Image Surveillance Assistant (ISA), reduces
the operator burdens of info overload and fatigue by culling
from the input stream input that is of no interest. We describe

• the full architecture,
• the current implementation, and
• directions for future extensions (i..e., more expressive
  semantic and perceptual representations and processing,
  and user interaction processes).

Read more here: https://sites.cc.gatech.edu/~alanwags/DLAI2016/(Maynord+)%20IJCAI-16%20DLAI%20WS.pdf