PointForward: Feedforward Driving Reconstruction through Point-Aligned Representations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chi, Cheng, Wang, Xianqi, Luo, Hongcheng, Tu, Mingfei, Xu, Gangwei, Zhang, Zehan, Wang, Bing, Chen, Guang, Ye, Hangjun, Peng, Sida, Yang, Xin, Sun, Haiyang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910211259760640
author Chi, Cheng
Wang, Xianqi
Luo, Hongcheng
Tu, Mingfei
Xu, Gangwei
Zhang, Zehan
Wang, Bing
Chen, Guang
Ye, Hangjun
Peng, Sida
Yang, Xin
Sun, Haiyang
author_facet Chi, Cheng
Wang, Xianqi
Luo, Hongcheng
Tu, Mingfei
Xu, Gangwei
Zhang, Zehan
Wang, Bing
Chen, Guang
Ye, Hangjun
Peng, Sida
Yang, Xin
Sun, Haiyang
contents High-fidelity reconstruction of driving scenes is crucial for autonomous driving. While recent feedforward 3D Gaussian Splatting (3DGS) methods enable fast reconstruction, their per-pixel Gaussian prediction paradigm often suffers from multi-view inconsistency and layering artifacts. Moreover, existing methods often model dynamic instances via dense flow prediction, which lacks explicit cross-view correspondence and instance-level consistency. In this paper, we propose PointForward, a feedforward driving reconstruction framework through point-aligned representations. Unlike pixel-aligned methods, we initialize sparse 3D queries in world space and aggregate multi-view image information via spatial-temporal fusion onto these queries, enforcing explicit cross-view consistency in a single feedforward pass. To handle scene dynamics, we introduce scene graphs that explicitly organize moving instances during reconstruction. By leveraging 3D bounding boxes, our method enables instance-level motion propagation and temporally consistent dynamic representations. Extensive experiments demonstrate that PointForward achieves state-of-the-art performance on large-scale driving benchmarks. The code will be available upon the publication of the paper.
format Preprint
id arxiv_https___arxiv_org_abs_2605_11594
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PointForward: Feedforward Driving Reconstruction through Point-Aligned Representations
Chi, Cheng
Wang, Xianqi
Luo, Hongcheng
Tu, Mingfei
Xu, Gangwei
Zhang, Zehan
Wang, Bing
Chen, Guang
Ye, Hangjun
Peng, Sida
Yang, Xin
Sun, Haiyang
Computer Vision and Pattern Recognition
High-fidelity reconstruction of driving scenes is crucial for autonomous driving. While recent feedforward 3D Gaussian Splatting (3DGS) methods enable fast reconstruction, their per-pixel Gaussian prediction paradigm often suffers from multi-view inconsistency and layering artifacts. Moreover, existing methods often model dynamic instances via dense flow prediction, which lacks explicit cross-view correspondence and instance-level consistency. In this paper, we propose PointForward, a feedforward driving reconstruction framework through point-aligned representations. Unlike pixel-aligned methods, we initialize sparse 3D queries in world space and aggregate multi-view image information via spatial-temporal fusion onto these queries, enforcing explicit cross-view consistency in a single feedforward pass. To handle scene dynamics, we introduce scene graphs that explicitly organize moving instances during reconstruction. By leveraging 3D bounding boxes, our method enables instance-level motion propagation and temporally consistent dynamic representations. Extensive experiments demonstrate that PointForward achieves state-of-the-art performance on large-scale driving benchmarks. The code will be available upon the publication of the paper.
title PointForward: Feedforward Driving Reconstruction through Point-Aligned Representations
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.11594