CP-Cert certifies outlier-robust pose registration up to 1000x faster than direct SDP solvers
zhenjun_zhao · x · 2026-09-05
From Timothy Barfoot's group, CP-Cert is a certifiable method for non-convex robotics problems whose SDP relaxations are degenerate. Starting from a candidate solution, it perturbs the primal solution into the interior of the PSD cone and converges back along the central path to obtain a global-optimality certificate efficiently via indirect linear algebra, sparsity, and parallelism. Applied to matrix-weighted pose registration and pointcloud data association (with a novel SDP relaxation), it achieves runtimes up to three orders of magnitude faster than state-of-the-art direct solvers.
More from Research
- LLMs exploiting Lean bugs is a short-term problem, author argues — avt_im · 2026-09-05
- FinFIRST benchmark tests agents on real financial research: source discovery, evidence selection and math — alifcoder · 2026-09-05
- Valeo.ai Brings 5 ECCV 2026 Papers on Driving Video Prediction and LVLM Safety — abursuc · 2026-09-05
- Sakana AI's Percept-Lens: a simple rule on frozen vision features detects AI images — SakanaAILabs · 2026-09-05
- VI3 anchors pretrained 3D foundation models to metric scale using only IMU readings — zhenjun_zhao · 2026-09-05
- ZipMVS compresses cost volumes for memory-efficient multi-view stereo reconstruction — zhenjun_zhao · 2026-09-05