Pith. sign in

REVIEW 10 cited by

DiST-4D: Disentangled Spatiotemporal Diffusion with Metric Depth for 4D Driving Scene Generation

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

arxiv 2503.15208 v1 pith:HS6GJMBN submitted 2025-03-19 cs.CV

classification cs.CV
keywords depthdist-4dmetricspatialdiffusiondrivinggeometricrepresentation
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Current generative models struggle to synthesize dynamic 4D driving scenes that simultaneously support temporal extrapolation and spatial novel view synthesis (NVS) without per-scene optimization. A key challenge lies in finding an efficient and generalizable geometric representation that seamlessly connects temporal and spatial synthesis. To address this, we propose DiST-4D, the first disentangled spatiotemporal diffusion framework for 4D driving scene generation, which leverages metric depth as the core geometric representation. DiST-4D decomposes the problem into two diffusion processes: DiST-T, which predicts future metric depth and multi-view RGB sequences directly from past observations, and DiST-S, which enables spatial NVS by training only on existing viewpoints while enforcing cycle consistency. This cycle consistency mechanism introduces a forward-backward rendering constraint, reducing the generalization gap between observed and unseen viewpoints. Metric depth is essential for both accurate reliable forecasting and accurate spatial NVS, as it provides a view-consistent geometric representation that generalizes well to unseen perspectives. Experiments demonstrate that DiST-4D achieves state-of-the-art performance in both temporal prediction and NVS tasks, while also delivering competitive performance in planning-related evaluations.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 10 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Point as Skeleton: Accumulated Point Cloud Enhanced Autoregressive Generation for Closed-Loop Autonomous Driving Simulation

    cs.CV 2026-07 conditional novelty 6.0 of 10

    Point-cloud skeleton conditions and a Reset-and-Roll inference scheme enable stable frame-wise autoregressive driving video generation for closed-loop autonomous driving simulation.

  2. DVGT: Driving Visual Geometry Transformer

    cs.CV 2025-12 conditional novelty 6.0 of 10

    DVGT predicts metric-scaled global 3D point maps and ego poses from unposed multi-view driving video, beating prior geometry models on several driving benchmarks.

  3. GeoWorld: Providing Full-frame Geometry Features to Facilitate 3D Scene Generation

    cs.CV 2025-11 conditional novelty 6.0 of 10

    GeoWorld improves image-to-3D scene generation by conditioning a video-diffusion model on full-frame geometry features extracted by a multi-view geometry model, yielding higher PSNR/SSIM/LPIPS than prior methods.

  4. Zero-Human Demonstration End-to-end Autonomous Driving with Trajectory Scorer

    cs.RO 2025-10 conditional novelty 6.0 of 10

    A reward-only offline RL method for trajectory planning in end-to-end autonomous driving achieves state-of-the-art on Navhard and competitive closed-loop HUGSIM performance without imitation learning.

  5. OmniNWM: Omniscient Driving Navigation World Models

    cs.CV 2025-10 conditional novelty 6.0 of 10

    OmniNWM jointly generates long panoramic multi-modal driving videos, controls them precisely via normalized Plücker ray-maps, and derives dense driving rewards from generated 3D occupancy.

  6. GS-Occ3D: Scaling Vision-only Occupancy Reconstruction with Gaussian Splatting

    cs.CV 2025-07 conditional novelty 6.0 of 10

    A camera-only Gaussian-surfel pipeline reconstructs full Waymo scenes, converts them to binary occupancy labels, and trains CVT-Occ to generalize on Occ3D-Waymo and Occ3D-nuScenes at a level close to or above LiDAR-la...

  7. ArbiViewGen: Controllable Arbitrary Viewpoint Camera Data Generation for Autonomous Driving via Stable Diffusion Models

    cs.CV 2025-08 conditional novelty 5.0 of 10

    ArbiViewGen generates arbitrary-viewpoint driving camera images by stitching the six input views into pseudo-target views and training a Stable Diffusion model to reconstruct the original views, enabling self-supervis...

  8. CRUISE: Cooperative Reconstruction and Editing in V2X Scenarios using Gaussian Splatting

    cs.CV 2025-07 conditional novelty 5.0 of 10

    CRUISE reconstructs real V2X driving scenes as editable Gaussians, then shows that training on its generated data improves 3D detection and tracking on the V2X-Seq benchmark.

  9. LongDWM: Cross-Granularity Distillation for Building a Long-Term Driving World Model

    cs.CV 2025-06 reject novelty 5.0 of 10

    A hierarchical coarse-to-fine diffusion transformer with cross-granularity distillation improves long-term driving video prediction, but the reported gains may be inflated by future-derived text prompts and a selected...

  10. Challenger: Affordable Adversarial Driving Video Generation

    cs.CV 2025-05 conditional novelty 5.0 of 10

    A framework for automatic generation of photorealistic adversarial driving videos, shown to sharply increase collision rates of end-to-end autonomous driving models.

Pith tools