Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

ShapeFit: Exact location recovery from corrupted pairwise directions

About

Let $t_1,\ldots,t_n \in \mathbb{R}^d$ and consider the location recovery problem: given a subset of pairwise direction observations $\{(t_i - t_j) / \|t_i - t_j\|_2\}_{i<j \in [n] \times [n]}$, where a constant fraction of these observations are arbitrarily corrupted, find $\{t_i\}_{i=1}^n$ up to a global translation and scale. We propose a novel algorithm for the location recovery problem, which consists of a simple convex program over $dn$ real variables. We prove that this program recovers a set of $n$ i.i.d. Gaussian locations exactly and with high probability if the observations are given by an \erdosrenyi graph, $d$ is large enough, and provided that at most a constant fraction of observations involving any particular location are adversarially corrupted. We also prove that the program exactly recovers Gaussian locations for $d=3$ if the fraction of corrupted observations at each location is, up to poly-logarithmic factors, at most a constant. Both of these recovery theorems are based on a set of deterministic conditions that we prove are sufficient for exact recovery.

Paul Hand, Choongbum Lee, Vladislav Voroninski• 2015

Related benchmarks

TaskDatasetResultRank
Translation Averaging1DSfM Montreal Notre Dame
Mean Translation Error0.96
14
Translation AveragingETH3D full coverage (test)
Mean Translation Error (t-bar) - Courtyard0.44
7
Translation Averaging1DSfM Alamo
Mean Translation Error1.78
7
Global Structure from MotionETH3D full coverage
Courtyard0.11
7
Translation Averaging1DSfM Piazza del Popolo
Mean Translation Error5.89
7
Runtime analysis1DSfM
Latency (ms)4.14
7
Translation Averaging1DSfM Tower of London
Mean Translation Error (t¯)16.8
7
Translation Averaging1DSfM NYC Library
Mean Translation Error14.27
7
Translation Averaging1DSfM Roman Forum
Mean Translation Error39.52
7
Translation Averaging1DSfM Vienna Cathedral
Mean Translation Error36.72
7
Showing 10 of 18 rows

Other info

Follow for update