B. Sabata et Jk. Aggarwal, SURFACE CORRESPONDENCE AND MOTION COMPUTATION FROM A PAIR OF RANGE IMAGES, Computer vision and image understanding, 63(2), 1996, pp. 232-250
Citations number
29
Categorie Soggetti
Computer Sciences, Special Topics","Computer Science Software Graphycs Programming
The estimation of the motion transformation of a moving object from a
sequence of images is of prime interest in computer vision. In this pa
per, the issues in estimating the motion parameters from a pair of ran
ge images are addressed, The motion estimation task, in the domain of
range image sequences, has two components: (1) extract the surfaces an
d establish the correspondence of the surfaces over the frames in the
sequence of range images, and (2) compute the motion transformation us
ing these surface correspondences. A novel procedure based on a hyperg
raph representation is presented for finding surface correspondence, T
wo scenes are modeled as hypergraphs and the hyperedges are matched us
ing a subgraph isomorphism algorithm. The hierarchical representation
of hypergraphs not only reduces the search space significantly but als
o facilitates the encoding of the topological and geometrical informat
ion used to direct the search procedure, Results obtained from real ra
nge image pairs show that the algorithm is robust and performs well in
presence of occlusions and incorrect segmentations. Motion transforma
tion between image frames is computed using the planar and the quadric
surface pairings, A least-squares minimization procedure is formulate
d that estimates the best motion transform, subject to the constraints
of rigid motion. For the case of linear feature pairings, the motion
computation becomes tractable because the rotation and the translation
computations become independent of each other. However, for quadric s
urfaces this is not true. The equation to be minimized is highly nonli
near and the uniqueness of solution cannot be guaranteed. The solution
obtained computes the motion by extracting unique linear features fro
m the quadric surfaces and using them to compute the motion transforma
tion, The main contribution of the work is a surface-based framework f
or motion estimation from a sequence of range images. The primary issu
es of correspondence and motion computation are formulated and solved
in terms of the surface descriptions. (C) 1996 Academic Press, Inc.