×

Scaleable object recognition with a belief model

  • US 7,184,994 B2
  • Filed: 06/18/2001
  • Issued: 02/27/2007
  • Est. Priority Date: 06/16/2000
  • Status: Expired due to Term
First Claim
Patent Images

1. A computer readable medium encoded with instructions for causing a computer to recognize objects in content by following the instructions to implement:

  • a blackboard comprisinga plurality of experts adapted to process data on a computer, and said data comprising at least one of original input data and/or data created by processing of any of said plurality of experts, anda controller operative to control said experts;

    a belief model, coupled to said controller, comprising a set of beliefs and probabilities associated with each belief of said set of beliefs, wherein said belief model comprises a set of rules deduced from a computer-implemented learning system, said learning system comprising truth data files for deducing said set of beliefs, said probabilities and shadow objects, a learning system controller and a statistics space controlled by said learning system controller, wherein said set of rules describes how different classes recognized by said learning system are related to each other spatially and physically;

    a belief network, coupled to said controller; and

    a relations subsystem, coupled to said controller.

View all claims
  • 8 Assignments
Timeline View
Assignment View
    ×
    ×