3D object recognition and pose with relational indexing

This paper addresses the problem of recognizing 3D objects from 2D intensit y images. It describes the object recognition system named RIO (relational indexing of objects), which contains a number of new techniques. RIO begins with an edge image obtained from a pair of intensity images taken with a s ingle camera and two different lightings. From the edge image, a set of new high-level features and relationships are extracted. and a technique calle d relational indexing is used to efficiently recall 2D view-class object mo dels that have similar relational descriptions from a potentially large dat abase of models. Once a model has been hypothesized, pairs of 2D-3D corresp onding features, including point pairs, line-segment pairs, and ellipse-cir cle pairs, are used in a new linear pose estimation framework to produce a hypothesized transformation from a 3D mesh model of the object to the image . The transformation is either accepted or rejected by a verification proce dure that projects the 3D model wireframe to the image and computes a Hausd orff-like distance measure between the projected model and the edge image. The resultant object recognition system is able to recognize 3D objects hav ing planar, cylindrical, and threaded surfaces in complex, multiobject scen es. (C) 2000 Academic Press.