TY  - JOUR
T1  - Certifiable relative pose estimation
A1  - Garcia-Salguero, Mercedes
A1  - Briales, Jesus
A1  - Gonzalez-Jimenez, Javier
JA  - Image and Vision Computing
Y1  - 2021
IS  - 63
SN  - 0262-8856
UR  - https://arxiv.org/pdf/2003.13732.pdf
M2  - doi: https://doi.org/10.1016/j.imavis.2021.104142
KW  - Certifiable algorithm
KW  - Convex programming
KW  - Epipolar constraint
KW  - Essential matrix
KW  - Linear Independence constraint qualification
KW  - Relative pose
N2  - In this paper we present the first fast optimality certifier for the non-minimal version of the Relative Pose problem for calibrated cameras from epipolar constraints. The proposed certifier is based on Lagrangian duality and relies on a novel closed-form expression for dual points. We also leverage an efficient solver that performs local optimization on the manifold of the original problem's non-convex domain. The optimality of the solution is then checked via our novel fast certifier. The extensive conducted experiments demonstrate that, despite its simplicity, this certifiable solver performs excellently on synthetic data, repeatedly attaining the (certified a posteriori) optimal solution and shows a satisfactory performance on real data.
M1  - img_url=http%3A%2F%2Fmapir.isa.uma.es%2Fimagesrepo%2Fpapers%2F2021%2F2021_mercedes_IMAVIS_cert_rpp_matlab.png
M1  - rank_indexname=
M1  - rank_pos_in_category=22
M1  - rank_num_in_category=110
M1  - rank_cat_name=Computer%20Science
M1  - impact_factor=3.860
ER  -