SynFlow: Scaling Up LiDAR Scene Flow Estimation with Synthetic Data

TL;DR

SynFlow introduces a synthetic LiDAR scene flow dataset generated through a motion-focused simulation pipeline, enabling models to learn robust motion priors that generalize well to real-world data, even with minimal real labels.

Contribution

The paper presents SynFlow, a large-scale synthetic dataset for LiDAR scene flow, and demonstrates its effectiveness in training models that generalize across real-world benchmarks.

Findings

Models trained on SynFlow-4k generalize well to real data in zero-shot settings.

Fine-tuning with 5% of real labels on SynFlow-4k surpasses models trained on full real data.

SynFlow-4k achieves a 34x scale-up over existing real-world benchmarks.

Abstract

Reliable 3D dynamic perception requires models that can anticipate motion beyond predefined categories, yet progress is hindered by the scarcity of dense, high-quality motion annotations. While self-supervision on unlabeled real data offers a path forward, empirical evidence suggests that scaling unlabeled data fails to close the performance gap due to noisy proxy signals. In this paper, we propose a shift in paradigm: learning robust real-world motion priors entirely from scalable simulation. We introduce SynFlow, a data generation pipeline that generates large-scale synthetic dataset specifically designed for LiDAR scene flow. Unlike prior works that prioritize sensor-specific realism, SynFlow employs a motion-oriented strategy to synthesize diverse kinematic patterns across 4,000 sequences ($\sim$940k frames), termed SynFlow-4k. This represents a 34x scale-up in annotated volume over existing real-world benchmarks. Our experiments demonstrate that SynFlow-4k provides a highly domain-invariant motion prior. In a zero-shot regime, models trained exclusively on our synthetic data generalize across multiple real-world benchmarks, rivaling in-domain supervised baselines on nuScenes and outperforming state-of-the-art methods on TruckScenes by 31.8%. Furthermore, SynFlow-4k serves as a label-efficient foundation: fine-tuning with only 5% of real-world labels surpasses models trained from scratch on the full available budget. We open-source the pipeline and dataset to facilitate research in generalizable 3D motion estimation. More detail can be found at https://kin-zhang.github.io/SynFlow.