MADrive: Memory-Augmented Driving Scene Modeling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Karpikova, Polina, Selikhanovych, Daniil, Struminsky, Kirill, Musaev, Ruslan, Golitsyna, Maria, Baranchuk, Dmitry
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917140227948544
author Karpikova, Polina
Selikhanovych, Daniil
Struminsky, Kirill
Musaev, Ruslan
Golitsyna, Maria
Baranchuk, Dmitry
author_facet Karpikova, Polina
Selikhanovych, Daniil
Struminsky, Kirill
Musaev, Ruslan
Golitsyna, Maria
Baranchuk, Dmitry
contents Recent advances in scene reconstruction have pushed toward highly realistic modeling of autonomous driving (AD) environments using 3D Gaussian splatting. However, the resulting reconstructions remain closely tied to the original observations and struggle to support photorealistic synthesis of significantly altered or novel driving scenarios. This work introduces MADrive, a memory-augmented reconstruction framework designed to extend the capabilities of existing scene reconstruction methods by replacing observed vehicles with visually similar 3D assets retrieved from a large-scale external memory bank. Specifically, we release MAD-Cars, a curated dataset of ${\sim}70$K 360° car videos captured in the wild and present a retrieval module that finds the most similar car instances in the memory bank, reconstructs the corresponding 3D assets from video, and integrates them into the target scene through orientation alignment and relighting. The resulting replacements provide complete multi-view representations of vehicles in the scene, enabling photorealistic synthesis of substantially altered configurations, as demonstrated in our experiments. Project page: https://yandex-research.github.io/madrive/
format Preprint
id arxiv_https___arxiv_org_abs_2506_21520
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MADrive: Memory-Augmented Driving Scene Modeling
Karpikova, Polina
Selikhanovych, Daniil
Struminsky, Kirill
Musaev, Ruslan
Golitsyna, Maria
Baranchuk, Dmitry
Computer Vision and Pattern Recognition
Recent advances in scene reconstruction have pushed toward highly realistic modeling of autonomous driving (AD) environments using 3D Gaussian splatting. However, the resulting reconstructions remain closely tied to the original observations and struggle to support photorealistic synthesis of significantly altered or novel driving scenarios. This work introduces MADrive, a memory-augmented reconstruction framework designed to extend the capabilities of existing scene reconstruction methods by replacing observed vehicles with visually similar 3D assets retrieved from a large-scale external memory bank. Specifically, we release MAD-Cars, a curated dataset of ${\sim}70$K 360° car videos captured in the wild and present a retrieval module that finds the most similar car instances in the memory bank, reconstructs the corresponding 3D assets from video, and integrates them into the target scene through orientation alignment and relighting. The resulting replacements provide complete multi-view representations of vehicles in the scene, enabling photorealistic synthesis of substantially altered configurations, as demonstrated in our experiments. Project page: https://yandex-research.github.io/madrive/
title MADrive: Memory-Augmented Driving Scene Modeling
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.21520