Model-Based Joint Motion and Structure Estimation from Stereo Images

摘要

In this paper we integrate motion and structure estimation in order to exploit their coherence. We propose an algorithm that uses models for object surfaces and their motion and estimates the model parameters using the image intensity matching criterion. The visible scene surface is represented with a parametrically deformable, spatially adaptive, wireframe model. Object motion is first modeled using the well-known rigid motion assumption along with the quaternion rotation representation. Nonrigid motion modeling using the finite element technique is also investigated as an alternative to rigid motion modeling or as a refinement of it. A functional containing image matching and surface smoothness constraints is minimized with respect to the unknown model parameters. A discontinuity detection scheme allowing deactivation of smoothness constraints across object boundaries, is investigated. A regularization scheme using a coarse to fine strategy is employed.