Object recognition entails identifying instances of known objects in sensory data by searching for a match between features in a scene and features on a model. The key elements that make object recognition feasible are the use of diverse sensory input forms such as stereo imagery or range data, appropriate low level processing of the sensory input, clever object representations, and good algorithms for scene-to-model hypothesis generation and model matching. Whether data acquisition takes place using video images or range sensors, an object recognition system must pre-process the sensory data for the extraction of relevant features in the scene. Once a feature vector is obtained, the problem now is that of correspondence. Provided a training session has taken place, a search for a match between model features and scene features is performed. A consistent match and the corresponding transformation give a solution to the problem of object recognition.

Peer Reviewed