mlsanityData to 7 Oct 2026

DrivingForward: Feed-forward 3D Gaussian Splatting for Driving Scene Reconstruction from Flexible Surround-view Input

arXiv 2409.12753 ↗First posted 19 Sep 2024cs.CVcited by 53

Not on the map. Why?

Abstract (arXiv)

We propose DrivingForward, a feed-forward Gaussian Splatting model that reconstructs driving scenes from flexible surround-view input. Driving scene images from vehicle-mounted cameras are typically sparse, with limited overlap, and the movement of the vehicle further complicates the acquisition of camera extrinsics. To tackle these challenges and achieve real-time reconstruction, we jointly train a pose network, a depth network, and a Gaussian network to predict the Gaussian primitives that represent the driving scenes. The pose network and depth network determine the position of the Gaussian primitives in a self-supervised manner, without using depth ground truth and camera extrinsics during training. The Gaussian network independently predicts primitive parameters from each input image, including covariance, opacity, and spherical harmonics coefficients. At the inference stage, our model can achieve feed-forward reconstruction from flexible multi-frame surround-view input. Experiments on the nuScenes dataset show that our model outperforms existing state-of-the-art feed-forward and scene-optimized reconstruction methods in terms of reconstruction.

Key facts. Cited by 53 papers in the map (data through 7 Oct 2026). Builds on 26 of its 38 references that are in the map. First posted on arXiv 19 Sep 2024.

Builds on

26 of 38 references.

References · 26

  1. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
  3. 2020pixelNeRF: Neural Radiance Fields from One or Few Imagescited by 1,574
  4. 2015Spatial Transformer Networkscited by 1,423
  5. 2022BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformerscited by 1,213
  6. 2022BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representationcited by 844
  7. 2016Unsupervised Monocular Depth Estimation with Left-Right Consistencycited by 819
  8. 2023pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstructioncited by 668
  9. 2023DrivingGaussian: Composite Gaussian Splatting for Surrounding Dynamic Autonomous Driving Scenescited by 359
  10. 2023Splatter Image: Ultra-Fast Single-View 3D Reconstructioncited by 359
  11. 2022BEVFusion: A Simple and Robust LiDAR-Camera Fusion Frameworkcited by 334
  12. 2024DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalizationcited by 256
14 more
  1. 2024Street Gaussians: Modeling Dynamic Urban Scenes with Gaussian Splattingcited by 218
  2. 2023EmerNeRF: Emergent Spatial-Temporal Scene Decomposition via Self-Supervisioncited by 216
  3. 2022FUTR3D: A Unified Sensor Fusion Framework for 3D Detectioncited by 194
  4. 2023SelfOcc: Self-Supervised Vision-Based 3D Occupancy Predictioncited by 133
  5. 2023StreetSurf: Extending Multi-view Implicit Surface Reconstruction to Street Viewscited by 120
  6. 2023Learning to Render Novel Views from Wide-Baseline Stereo Pairscited by 103
  7. 2023COTR: Compact Occupancy TRansformer for Vision-based 3D Occupancy Predictioncited by 94
  8. 2022SurroundDepth: Entangling Surrounding Views for Self-Supervised Multi-Camera Depth Estimationcited by 76
  9. 2023UniPAD: A Universal Pre-training Paradigm for Autonomous Drivingcited by 73
  10. 2024AutoSplat: Constrained Gaussian Splatting for Autonomous Driving Scene Reconstructioncited by 52
  11. 2021Full Surround Monodepth from Multiple Camerascited by 41
  12. 2024FastLGS: Speeding up Language Embedded Gaussians with Feature Grid Mappingcited by 34
  13. 2024DistillNeRF: Perceiving 3D Scenes from Single-Glance Images by Distilling Neural Fields and Foundation Model Featurescited by 33
  14. 2022Image Understands Point Cloud: Weakly Supervised 3D Semantic Segmentation via Association Learningcited by 15

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 3; 2025: 29; 2026: 21.

53 papers.

Sort:

By area:

  1. 2025Cosmos-Drive-Dreams: Scalable Synthetic Driving Data Generation with World Foundation Modelscited by 80
  2. 2025DiST-4D: Disentangled Spatiotemporal Diffusion with Metric Depth for 4D Driving Scene Generationcited by 39
  3. 2024DrivingRecon: Large 4D Gaussian Reconstruction Model For Autonomous Drivingcited by 27
  4. 2025GaussianDWM: 3D Gaussian Driving World Model for Unified Scene Understanding and Multi-Modal Generationcited by 23
  5. 2025EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesiscited by 20
  6. 2024PixelGaussian: Generalizable 3D Gaussian Reconstruction from Arbitrary Viewscited by 18
  7. 2025PartRM: Modeling Part-Level Dynamics with Large Cross-State Reconstruction Modelcited by 15
  8. 2024Bench2Drive-R: Turning Real World Data into Reactive Closed-Loop Autonomous Driving Benchmark by Generative Modelcited by 14
  9. 2025AutoOcc: Automatic Open-Ended Semantic Occupancy Annotation via Vision-Language Guided Gaussian Splattingcited by 14
  10. 2025DriveGen3D: Boosting Feed-Forward Driving Scene Generation with Efficient Video Diffusioncited by 14
  11. 2025UniForward: Unified 3D Scene and Semantic Field Reconstruction via Feed-Forward Gaussian Splatting from Only Sparse-View Imagescited by 13
  12. 2025NaviMaster: Learning a Unified Policy for GUI and Embodied Navigation Taskscited by 10
38 more
  1. 2025Learning-based 3D Reconstruction in Autonomous Driving: A Comprehensive Surveycited by 9
  2. 2025MuDG: Taming Multi-modal Diffusion with Gaussian Splatting for Urban Scene Reconstructioncited by 8
  3. 20254D Driving Scene Generation With Stereo Forcingcited by 8
  4. 2025UniSplat: Unified Spatio-Temporal Fusion via 3D Latent Scaffolds for Dynamic Driving Scene Reconstructioncited by 8
  5. 2025VGD: Visual Geometry Gaussian Splatting for Feed-Forward Surround-view Driving Reconstructioncited by 7
  6. 2025VR-Drive: Viewpoint-Robust End-to-End Driving with Feed-Forward 3D Gaussian Splattingcited by 5
  7. 2025OmniGen: Unified Multimodal Sensor Generation for Autonomous Drivingcited by 5
  8. 2026ReconDrive: Fast Feed-Forward 4D Gaussian Splatting for Autonomous Driving Scene Reconstructioncited by 5
  9. 2025Progressive Gaussian Transformer with Anisotropy-aware Sampling for Open Vocabulary Occupancy Predictioncited by 4
  10. 2025FMLGS: Fast Multilevel Language Embedded Gaussians for Part-level Interactive Agentscited by 3
  11. 2025Seeing Clearly, Forgetting Deeply: Revisiting Fine-Tuned Video Generators for Driving Simulationcited by 3
  12. 2025SparseWorld-TC: Trajectory-Conditioned Sparse Occupancy World Modelcited by 3
  13. 2026EVolSplat4D: Efficient Volume-based Gaussian Splatting for 4D Urban Scene Synthesiscited by 3
  14. 2026UFO: Unifying Feed-Forward and Optimization-based Methods for Large Driving Scene Modelingcited by 3
  15. 2025Learning Unified Representation of 3D Gaussian Splattingcited by 2
  16. 2025DrivingScene: A Multi-Task Online Feed-Forward 3D Gaussian Splatting Method for Dynamic Driving Scenescited by 2
  17. 2025SRSplat: Feed-Forward Super-Resolution Gaussian Splatting from Sparse Multi-View Imagescited by 2
  18. 2026Learning to Drive is a Free Gift: Large-Scale Label-Free Autonomy Pretraining from Unposed In-The-Wild Videoscited by 2
  19. 2026Ground4D: Spatially-Grounded Feedforward 4D Reconstruction for Unstructured Off-Road Scenescited by 2
  20. 2026Instant NuRec: Feed-Forward 3D Gaussian Reconstruction for Driving Scene Simulationcited by 2
  21. 2025sshELF: Single-Shot Hierarchical Extrapolation of Latent Features for 3D Reconstruction from Sparse-Viewscited by 1
  22. 2025LidarPainter: One-Step Away From Any Lidar View To Novel Guidancecited by 1
  23. 2025PFDepth: Heterogeneous Pinhole-Fisheye Joint Depth Estimation via Distortion-aware Gaussian-Splatted Volumetric Fusioncited by 1
  24. 2025XYZCylinder: Towards Compatible Feed-Forward 3D Gaussian Splatting for Driving Scenes via Unified Cylinder Lifting Methodcited by 1
  25. 2026PointForward: Feedforward Driving Reconstruction through Point-Aligned Representationscited by 1
  26. 2026SurroundNEXO: Ego-Centric Metric Bridging for Spatially Consistent Geometry in Autonomous Drivingcited by 1
  27. 2026Bridging 3D Gaussians and Semantic Occupancy for Comprehensive Open-Vocabulary Scene Understanding from Unposed Imagescited by 1
  28. 2025ADGaussian: Generalizable Gaussian Splatting for Autonomous Driving via Multi-modal Joint Learningcited by 0
  29. 2025FreeGen: Feed-Forward Reconstruction-Generation Co-Training for Free-Viewpoint Driving Scene Synthesiscited by 0
  30. 2026S2D: Sparse to Dense Lifting for 3D Reconstruction with Minimal Inputscited by 0
  31. 2026VisionNVS: Self-Supervised Inpainting for Novel View Synthesis under the Virtual-Shift Paradigmcited by 0
  32. 2026World2Minecraft: Occupancy-Driven Simulated Scenes Constructioncited by 0
  33. 2026ConFixGS: Learning to Fix Feedforward 3D Gaussian Splatting with Confidence-Aware Diffusion Priors in Driving Scenescited by 0
  34. 2026FRUC: Feedforward Dynamic Scene Reconstruction from Uncalibrated Collaborative Driving Viewscited by 0
  35. 2026DeblurNVS: Geometric Latent Diffusion for Novel View Synthesis from Sparse Motion-Blurred Imagescited by 0
  36. 2026DrivingDepth: Sparse-Prompted Pixel-wise Scale Correction for Driving Depth Estimationcited by 0
  37. 2026Targeted Structure Completion for Sparse-View 3D Reconstruction in Autonomous Drivingcited by 0
  38. 2026DerainSplat: Feed-Forward Clean 3D Gaussian Splatting from Sparse Rainy Viewscited by 0

50 most cited of 53 All 53 (JSON)