Using single perspective views for 3-D object recognition

A method to recognize 3-D planar-faced objects from single views, using the perspective projection model, is presented. Lines, junctions and faces are considered as the main 3-D object descriptors. These features do not depend on a particular point of view, but correspond to geometric descriptive properties of the object. The geometric constraints related to the perspective projection of these object features are analyzed, and exploited for shortening the efforts in a recognition process.<
>