3D Recognition and Pose Estimation Using Model Reconstructed by Pairwise Correspondence
2015-12-21 16:27:16   来源：   评论：0 点击：

Abstract—This paper focuses on pose estimation of 3D object from 2D images. The base system of Structure from Motion is applied. A 3D model with feature points and descriptors is prebuilt. 3D-2D matching is employed in recognition and pose estimation process. The method of pairwise correspondence is proposed. To reduce reprojection error of the model, a weighting step is added. Viewpoint change of camera is simulated by affine transformation in 3D model description. Simulated affine feature is modified to keep the model compact. In experiment section, the performance of the pose estimation system as well as the modified affine feature are tested, and the results are compared to related previous research.
 

Index Terms—pose estimation, 3D model, 3D recognition, affine feature

Cite: Lin Tian, "3D Recognition and Pose Estimation Using Model Reconstructed by Pairwise Correspondence," Lecture Notes on Information Theory, Vol. 3, No. 2, pp. 65-69, December 2015. doi: 10.18178/lnit.3.2.65-69