On the Tightness of Semidefinite Relaxations for Rotation Estimation

Abstract

Why is it that semidefinite relaxations have been so successful in numerous applications in computer vision and robotics for solving non-convex optimization problems involving rotations? In studying the empirical performance, we note that there are few failure cases reported in the literature, in particular for estimation problems with a single rotation, motivating us to gain further theoretical understanding. A general framework based on tools from algebraic geometry is introduced for analyzing the power of semidefinite relaxations of problems with quadratic objective functions and rotational constraints. Applications include registration, hand–eye calibration, and rotation averaging. We characterize the extreme points and show that there exist failure cases for which the relaxation is not tight, even in the case of a single rotation. We also show that some problem classes are always tight given an appropriate parametrization. Our theoretical findings are accompanied with numerical simulations, providing further evidence and understanding of the results.

Newton-PnP: Real-time Visual Navigation for Autonomous Toy-Drones

“…A main open problem is to remove Assumption 1. While our SOS solver always returned the optimal solution in practice as expected by [9], there is synthetic input where it should fail [3]. Since in practice the number of required Newton steps was roughly 15, we believe that the running time of our algorithm can be reduced to only 𝑛 + log(1/𝜀).…”

“…Nevertheless, recent studies showed that, in the case of computer vision applications, the SOS relaxation usually yields the optimal solution in practice. See possible explanations in [9] and many references therein. However, SOS method either provides a proof, called certificate, that the returned solution is 𝜀-optimal, or states that it failed.…”

A Fast and Reliable Solution to PnP, Using Polynomial Homogeneity and a Theorem of Hilbert

“…This immediately raises the following question: Are there PnP instances in which the solution of the fourth-degree problem is different from the correct one (i.e., the global minimum of )? To the best of our knowledge, this question was first answered (in the affirmative) in two recent papers [ 11 , 17 ]. These cases appear to be very rare, as described in [ 17 ].…”

Deep Learning for 3D Reconstruction, Augmentation, and Registration: A Review Paper

“…Lagrangian duality and SDP relaxations were used to tackle the multiple point cloud registration problem. This problem was investigated further in this model, where it was demonstrated that the SDP relaxation is always tight under low-noise regimes [ 151 ]. A study of global optimality requirements for point set registration (PSR) with incomplete data was presented using this approach.…”