Nonlinear dimensionality reduction is an active area of research. In this paper, we present a thematically different approach to detect low-dimensional manifold that lies within set bounds derived from given point cloud. A matrix representing distances on low-rank, and our method based current low-rank Matrix Completion (MC) techniques for recovering partially observed fully entries. MC methods...

Any non-complete orthonormal system in a Hilbert space can be transformed into basis by small perturbations.

Automatic assembly is a promising research topic in 3D computer vision and robotics. Existing works focus on generating (e.g., IKEA furniture) from scratch with set of parts, namely part assembly. In practice, there are higher demands for the robot to take over finish an incomplete half-assembled off-the-shelf toolkit, especially human-robot multi-agent collaborations. Compared assembly, it mor...