Mind's Eye
Abstract
The Mind's Eye program is developing a machine-based capability to learn generative representations of action between objects in a scene, directly from visual inputs, and then to reason over those learned representations. Mind's Eye will create the perceptual and cognitive underpinnings for reasoning about the action in scenes, enabling the generation of a narrative description of the action taking place in the visual field. The technologies developed under Mind's Eye have applicability in automated ground-based surveillance systems. This effort is funded in PE 0602702E, Project TT-13 in FY 2013.
Document Details
- Document Type
- Accomplishment
- Publication Date
- Oct 01, 2014
- Source ID
- 10deca5de5b406cd82f213ceca828d90
Related Documents
- Root: MACHINE INTELLIGENCE