paper-with-me

홈 › Papers

LiDAR Prompted Spatio-Temporal Multi-View Stereo for Autonomous Driving

2026-03-04 · Qihao Sun, Jiarun Liu, Ziqian Ni, Jianyun Xu, Tao Xie, Lijun Zhao, Ruifeng Li, Sheng Yang arxiv

Accurate metric depth is critical for autonomous driving perception and simulation, yet current approaches struggle to achieve high metric accuracy, multi-view and temporal consistency, and cross-domain generalization. To address these challenges, we present DriveMVS, a novel multi-view stereo framework that reconciles these competing objectives through two key insights: (1) Sparse but metrically accurate LiDAR observations can serve as geometric prompts to anchor depth estimation in absolute scale, and (2) deep fusion of diverse cues is essential for resolving ambiguities and enhancing robustness, while a spatio-temporal decoder ensures consistency across frames. Built upon these principles, DriveMVS embeds the LiDAR prompt in two ways: as a hard geometric prior that anchors the cost volume, and as soft feature-wise guidance fused by a triple-cue combiner. Regarding temporal consistency, DriveMVS employs a spatio-temporal decoder that jointly leverages geometric cues from the MVS cost volume and temporal context from neighboring frames. Experiments show that DriveMVS achieves state-of-the-art performance on multiple benchmarks, excelling in metric accuracy, temporal stability, and zero-shot cross-domain transfer, demonstrating its practical value for scalable, reliable autonomous driving systems.

📄 PDF Abstract BibTeX arXiv:2603.03765

Code (0)

등록된 구현이 없습니다.

Tasks

Domain GeneralizationAutonomous DrivingDepth Estimation

Similar Papers 제목 키워드 기반

Bootstrapping a 4D LiDAR Annotation Tool from Video Foundation Models

2026-08-26 · Jihun Kim, Hyun-Kurl Jang, Hyemin Yang, Jinnyeong Yang 외 arxiv

Progress in 4D LiDAR segmentation is bottlenecked by data. Assigning temporally consistent labels across sparse point cloud sequences is costly and hard to scale, and every new task or domain tends to demand fresh dense …

Scene UnderstandingVideo Segmentation

AnyView: Synthesizing Any Novel View in Dynamic Scenes

2026-01-23 · Basile Van Hoorick, Dian Chen, Shun Iwase, Pavel Tokmakov 외 arxiv

Modern generative video models excel at producing convincing, high-quality outputs, but struggle to maintain multi-view and spatiotemporal consistency in highly dynamic real-world environments. In this work, we introduce…

Video Generation

MVFuseNet: Improving End-to-End Object Detection and Motion Forecasting through Multi-View Fusion of LiDAR Data

2021-04-21 · Ankit Laddha, Shivam Gautam, Stefan Palombo, Shreyash Pandey 외

In this work, we propose \textit{MVFuseNet}, a novel end-to-end method for joint object detection and motion forecasting from a temporal sequence of LiDAR data. Most existing methods operate in a single view by projectin…

Motion Forecastingobject-detectionObject Detection

GSPR: Multimodal Place Recognition Using 3D Gaussian Splatting for Autonomous Driving

2024-10-01 · Zhangshuo Qi, Junyi Ma, Jingyi Xu, Zijie Zhou 외

Place recognition is a crucial module to ensure autonomous vehicles obtain usable localization information in GPS-denied environments. In recent years, multimodal place recognition methods have gained increasing attentio…

Autonomous DrivingAutonomous Vehicles

SuperFlow++: Enhanced Spatiotemporal Consistency for Cross-Modal Data Pretraining

2025-03-25 · Xiang Xu, Lingdong Kong, Hui Shuai, Wenwei Zhang 외

LiDAR representation learning has emerged as a promising approach to reducing reliance on costly and labor-intensive human annotations. While existing methods primarily focus on spatial alignment between LiDAR and camera…

Autonomous DrivingComputational EfficiencyContrastive LearningRepresentation Learning+1