The generalised Hough transform[13] is commonly used [11,14,15,16,17]. Such approaches involve an accumulation of evidence which either supports or weakens a hypothesis on the pose of an object in the image. Objects are recognised by means of a rotation, scaling, translation (RST) transformation which maps observed image features onto features in the model. This is achieved by assuming that the object in the image can be represented by an instance of the model having undergone some RST transformation and that the object is rigid. Even with such limitations the approach is viable. A match is still possible even if some model features are missing, and if spurious features are found. As a result the method is able to cope with both noisy and partially occluded objects. We present a method of recognising 3-D objects and determining their position and orientation from single 2-D images of both noisy and partially occluded shapes. Our approach combines analytical and structural-type techniques, adopting a 2-D pose-clustering method combined with a comparison to a database of 2-D model descriptions for various object orientations. Object features are used both to control the amount of computation and to direct the search of the database. Features derived from an analysis of the local geometry of the object boundary are used in matching. Local features are preferable to global techniques which perform poorly in noisy conditions and where the object boundary is partially occluded. In these circumstances, local features provide sufficient information to enable the object to be recognised and to determine orientation to within less than seven degrees of arc, on average, in both the major and minor axes of the object.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.