REVIEW 9 cited by
MTGS: Multi-Traversal Gaussian Splatting
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
MTGS: Multi-Traversal Gaussian Splatting
read the original abstract
Multi-traversal data, commonly collected through daily commutes or by self-driving fleets, provides multiple viewpoints for scene reconstruction within a road block. This data offers significant potential for high-quality novel view synthesis, which is crucial for applications such as autonomous vehicle simulators. However, inherent challenges in multi-traversal data often result in suboptimal reconstruction quality, including variations in appearance and the presence of dynamic objects. To address these issues, we propose Multi-Traversal Gaussian Splatting (MTGS), a novel approach that reconstructs high-quality driving scenes from arbitrarily collected multi-traversal data by modeling a shared static geometry while separately handling dynamic elements and appearance variations. Our method employs a multi-traversal dynamic scene graph with a shared static node and traversal-specific dynamic nodes, complemented by color correction nodes with learnable spherical harmonics coefficient residuals. This approach enables high-fidelity novel view synthesis and provides flexibility to navigate any viewpoint. We conduct extensive experiments on a large-scale driving dataset, nuPlan, with multi-traversal data. Our results demonstrate that MTGS improves LPIPS by 23.5% and geometry accuracy by 46.3% compared to single-traversal baselines. The code and data would be available to the public.
Forward citations
Cited by 9 Pith papers
-
Closed Loop Dynamic Driving Data Mixture for Real-Synthetic Co-Training
AutoScale is a closed-loop data engine using Graph-RAE for scene representation and Cluster-GA for importance-based retrieval to improve real-synthetic co-training for autonomous driving.
-
Appearance Decomposition Gaussian Splatting for Multi-Traversal Reconstruction
ADM-GS decomposes static background appearance into traversal-invariant material and traversal-dependent illumination via a frequency-separated neural light field, yielding +0.98 dB PSNR gains and better cross-travers...
-
World Engine: Towards the Era of Post-Training for Autonomous Driving
World Engine generates realistic safety-critical driving variations from logs for reinforcement post-training, reducing benchmark failures more than data scaling and showing collision reductions plus on-road gains in ...
-
P2GS: Physical Prior-guided Gaussian Splatting for Photometrically Consistent Urban Reconstruction
P2GS jointly decomposes LDR images into a view-invariant linear HDR radiance field, per-view exposure scales, and tone-mapping functions without HDR supervision to enforce photometric consistency in urban Gaussian Splatting.
-
LEAD: Minimizing Learner-Expert Asymmetry in End-to-End Driving
Reducing expert-student asymmetries in visibility, uncertainty, and route specification enables a new TransFuser v6 policy that reaches 95 DS on Bench2Drive and more than doubles prior scores on Longest6 v2 and Town13.
-
Optimization-Guided Diffusion for Interactive Scene Generation
OMEGA guides diffusion sampling with per-step constrained optimization and game-theoretic adversarial modeling to generate physically valid and interactive driving scenes, raising valid scene ratios from 32% to 72% an...
-
SimScale: Learning to Drive via Real-World Simulation at Scale
SimScale synthesizes unseen driving states from real logs via neural rendering and reactive environments, generates pseudo-expert trajectories, and shows that co-training on real plus simulated data improves planning ...
-
Towards Physically Consistent 4D Scene Reconstruction for Closed-loop Autonomous Driving Simulation
Introduces Orthogonal Projected Gradient (OPG) and a smoothness-based temporal regularization to restore spatial identifiability and ensure physically consistent 4D scene reconstruction for closed-loop autonomous driv...
-
Does it matter which Gaussians you pick in 4D Gaussian streaming?
A reinforcement-learned plug-in sampler can match or beat IGS@8192 quality on N3DV and MeetingRoom using as few as 256 anchors while reducing per-frame time.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.