Object Recognition Engine (ORE)

Recognition of objects of “known” types in video and static pictures, extensible framework for adding new object types (training).



Recognition of objects in video and static pictures

  • Highly adaptive. Robust to different backgrounds of image, image orientation and scale.
  • Very quick. Easily work with real-time applications for detection of objects on video.



Framework for "training" and building knowledge base

  • Extensible framework for recognition of "known" types of objects.
  • "Training" of new types of objects and adding them to object type knowledge base.
  • "Detalization" of known types in dividing into subtypes.



Support of most demanded scenarios:

  • SELECT ONE - out of finite number of known object types select one that is the closest match for the object being tested.
  • ONE OR NONE - determine if the object being tested belongs to any of the known types, and if it does, which one.
  • PATTERN - search of a definite image pattern, such as logo, trade mark, etc.