A Framework for Reconstruction based Recognition of Partially Occluded Repeated Objects

摘要: In this paper we propose a reconstruction based recognition scheme for objects with repeated components, using single image of such configuration, in which one the components may be partially occluded. our strategy reconstruct each respect to same frame and use these compute invariants.We new mathematical framework projective affinely objects. This uses repetition explicitly hence is able handle substantial occlusion components. We then apply pair quadrics. The information required are outline conic quadrics correspondence between any four points images general position on quadric its repetition. Projective invariants computed reconstructed have been used recognition. has applied monuments multi-dome architecture. Experiments established discriminatory ability invariants.