PCSTracker: Long-Term Scene Flow Estimation for Point Cloud Sequences

2026-03-20

AI Summary

Existing methods for long-term scene flow estimation on point cloud sequences struggle to maintain temporal consistency due to geometric changes, occlusions, and error accumulation over time. To address this challenge, this work proposes PCSTracker, the first end-to-end framework that achieves temporally coherent dynamic scene flow estimation through an iterative geometry-motion joint optimization (IGMO) module and a spatiotemporal point trajectory update (STTU) mechanism, integrated within an overlapping sliding window inference strategy. The proposed approach achieves state-of-the-art accuracy on both PointOdyssey3D and ADT3D benchmarks while running at 32.5 frames per second, significantly outperforming existing RGB-D-based methods in both precision and efficiency.

Technology Category

• Computer Vision: Motion & Tracking
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Point cloud scene flow estimation is fundamental to long-term and fine-grained 3D motion analysis. However, existing methods are typically limited to pairwise settings and struggle to maintain temporal consistency over long sequences as geometry evolves, occlusions emerge, and errors accumulate. In this work, we propose PCSTracker, the first end-to-end framework specifically designed for consistent scene flow estimation in point cloud sequences. Specifically, we introduce an iterative geometry motion joint optimization module (IGMO) that explicitly models the temporal evolution of point features to alleviate correspondence inconsistencies caused by dynamic geometric changes. In addition, a spatio-temporal point trajectory update module (STTU) is proposed to leverage broad temporal context to infer plausible positions for occluded points, ensuring coherent motion estimation. To further handle long sequences, we employ an overlapping sliding-window inference strategy that alternates cross-window propagation and in-window refinement, effectively suppressing error accumulation and maintaining stable long-term motion consistency. Extensive experiments on the synthetic PointOdyssey3D and real-world ADT3D datasets show that PCSTracker achieves the best accuracy in long-term scene flow estimation and maintains real-time performance at 32.5 FPS, while demonstrating superior 3D motion understanding compared to RGB-D-based approaches.

Problem

• scene flow estimation
• point cloud sequences
• temporal consistency
• long-term motion analysis
• occlusion handling

Innovation

• scene flow estimation
• point cloud sequences
• temporal consistency
• occlusion handling
• sliding-window inference