PolarMOT: How Far Can Geometric Relations Take Us in 3D Multi-Object Tracking?

TL;DR

PolarMOT demonstrates that encoding geometric relations alone, using a graph neural network, can achieve state-of-the-art results in 3D multi-object tracking, generalizing well across diverse datasets and locations.

Contribution

This work introduces a geometric-only approach for 3D multi-object tracking using graph neural networks, avoiding appearance cues and achieving high generalization.

Findings

Achieved state-of-the-art on nuScenes dataset.

Generalizes well across different cities and datasets.

Effective in scenes with non-holonomic motion.

Abstract

Most (3D) multi-object tracking methods rely on appearance-based cues for data association. By contrast, we investigate how far we can get by only encoding geometric relationships between objects in 3D space as cues for data-driven data association. We encode 3D detections as nodes in a graph, where spatial and temporal pairwise relations among objects are encoded via localized polar coordinates on graph edges. This representation makes our geometric relations invariant to global transformations and smooth trajectory changes, especially under non-holonomic motion. This allows our graph neural network to learn to effectively encode temporal and spatial interactions and fully leverage contextual and motion cues to obtain final scene interpretation by posing data association as edge classification. We establish a new state-of-the-art on nuScenes dataset and, more importantly, show that our method, PolarMOT, generalizes remarkably well across different locations (Boston, Singapore, Karlsruhe) and datasets (nuScenes and KITTI).