Voronoi-based Second-order Descriptor with Whitened Metric in LiDAR Place Recognition

TL;DR

This paper introduces a novel Voronoi-based second-order pooling method with whitening for LiDAR place recognition, improving descriptor robustness and Euclidean distance suitability.

Contribution

It integrates Voronoi cell bias with second-order pooling and whitening, addressing instability issues and enhancing LiDAR place recognition performance.

Findings

Improved accuracy on Oxford Robotcar benchmark.

Enhanced descriptor stability with whitening technique.

Better Euclidean distance compatibility in descriptors.

Abstract

The pooling layer plays a vital role in aggregating local descriptors into the metrizable global descriptor in the LiDAR Place Recognition (LPR). In particular, the second-order pooling is capable of capturing higher-order interactions among local descriptors. However, its existing methods in the LPR adhere to conventional implementations and post-normalization, and incur the descriptor unsuitable for Euclidean distancing. Based on the recent interpretation that associates NetVLAD with the second-order statistics, we propose to integrate second-order pooling with the inductive bias from Voronoi cells. Our novel pooling method aggregates local descriptors to form the second-order matrix and whitens the global descriptor to implicitly measure the Mahalanobis distance while conserving the cluster property from Voronoi cells, addressing its numerical instability during learning with diverse techniques. We demonstrate its performance gains through the experiments conducted on the Oxford Robotcar and Wild-Places benchmarks and analyze the numerical effect of the proposed whitening algorithm.