TrackCue is a self‑supervised framework that enhances LiDAR scene flow estimation by using dense image‑based point tracking to generate reliable static‑dynamic supervision. It anchors tracking queries to projected LiDAR points, applies a visually consistent motion compensation strategy to isolate true object motion, and lifts image‑space motion cues back to LiDAR points for refined pseudo‑labels. This approach reduces noise from sparse or occluded LiDAR geometry and improves the precision of dynamic pseudo‑labels, leading to better scene flow learning.
arXiv:2607.29237v2 Announce Type: replace Abstract: LiDAR scene flow is the task of estimating point-wise motion from sequential point clouds. It is essential for autonomous driving, as it provides c...
RVLoss introduces a runoff vote mechanism for self‑supervised LiDAR scene flow estimation, addressing motion rigidity by grouping nearest‑neighbor derived motions into dominant flow candidates and selecting the most consistent one through a two‑stage voting process. This approach generates cluster‑wise rigid flows and free‑form flows as pseudo‑labels, enabling seamless integration into existing feedforward architectures. Experiments on the Argoverse2 2026 Challenge demonstrate that models trained with RVLoss outperform baseline self‑supervised methods by 20% and maintain consistent gains across four additional datasets.
TAPVid-MV is a new benchmark for tracking any point in 3D across multiple synchronized camera views. It comprises 284 sequences, 1,142 calibrated camera streams, and 109,769 point tracks, covering indoor and outdoor domains and derived from various modalities such as depth, LiDAR, SLAM, and simulation. The dataset is visually verified, and evaluation shows that current multi‑view trackers do not consistently outperform monocular trackers, highlighting geometry recovery as a key bottleneck.
TrackEverything is a 3D point tracker that overcomes the trade‑off between sparse long‑horizon tracking and dense short‑clip tracking by representing videos as persistent 3D scene tracks in world coordinates. It introduces voxel‑based de‑duplication at sliding‑window boundaries, a two‑stage refinement process (endpoint refiner and lightweight trajectory refiner), and a 3D WAFT module that replaces memory‑heavy 4D correlation volumes with efficient feature sampling. The method can track all visible points in videos longer than 1000 frames using only 40 GB of GPU memory, outperforming existing dense trackers on short clips and matching sparse trackers on long sequences.
Scene coordinate regression (SCR) achieves strong performance in outdoor LiDAR localization, but it usually requires scene-specific training that can take days, limiting practical deployment. Recent w...
Synthesizing a novel-view video from a monocular reference video along a target camera trajectory requires both geometric consistency and motion fidelity with respect to the reference video. Existing methods based on explicit 3D representations are limited by the accuracy of off-the-shelf reconstruction modules, which often produce inaccurate geometry for dynamic objects in monocular videos.
The paper introduces LiDAR‑SAM2, a framework that converts the 2D video foundation model SAM2 into a scalable source of supervision for 4D LiDAR data. By projecting SAM2 video masks into multi‑view LiDAR space and aggregating them temporally, the method automatically generates temporally coherent LiDAR labels without human annotation. Experiments on SemanticKITTI show that these automatically produced semantic and panoptic labels achieve quality close to full human annotation, enabling models trained on them to approach the performance of fully supervised systems.
LightLoc++ introduces a sensor‑robust framework for outdoor LiDAR localization that decouples a pretrained backbone from lightweight scene‑specific heads, enabling efficient new‑scene training. The authors present SULID, a synchronized urban multi‑LiDAR dataset featuring 32‑, 64‑, and 128‑beam sensors, which is used to pretrain a backbone via cross‑sensor consistency learning. Additional techniques such as sample classification guidance and redundant sample downsampling reduce regression ambiguity and computational load, achieving state‑of‑the‑art performance with minimal new‑scene training cost.
MeshBind introduces an automatic annotation pipeline that generates per-point scene flow for non‑rigid objects such as pedestrians by reconstructing articulated mesh models from six surround cameras and registering them to LiDAR data. The method employs a local‑frame binding technique that attaches each point to a mesh via a per‑triangle tangent‑normal frame, and a three‑layer flow registration pipeline that compensates ego‑vehicle motion, performs box binding, and applies mesh binding to produce full‑scene flow. MeshBind is implemented in the SF‑Studio tool, and the authors release the SUScape‑Flow Benchmark (SFB), a large‑scale dataset with 41k frames of per‑point non‑rigid pedestrian annotations and a comprehensive evaluation protocol, demonstrating that fine‑tuning on SFB reduces endpoint error by 2‑3× and equalizes non‑rigid and rigid EPE.
arXiv:2506. 22784v2 Announce Type: replace-cross Abstract: Point-pixel registration between LiDAR point clouds and camera images is a fundamental yet challenging task in autonomous driving and robotic perception.
AutoExpert proposes a new benchmark for automating 3D LiDAR annotation using expert‑crafted guidelines, redefining evaluation protocols and re‑annotating nuScenes and PandaSet datasets. The pipeline combines 2D detection and segmentation in RGB images, lifts these detections into 3D with known sensor poses, and generates 3D cuboids, improving 3D detection mAP from 12.1 to 25.4 on the AutoExpert‑nuScenes benchmark.
SplatLabel is an automated pipeline that uses a 4D Gaussian representation to generate LiDAR segmentation and semantic occupancy grids with predictive confidence. It models dynamic scenes through an explicit temporal manifold, tracking moving actors without requiring pre‑annotated 3D bounding boxes. By integrating 360‑degree LiDAR depth maps and distilling soft probabilities from 2D models, it resolves semantic ambiguities over time and space, and evaluates pseudo‑labels via a selective classification framework that balances precision and recall.