MADrive: Memory-Augmented Driving Scene Modeling
This work addresses the challenge of generating realistic driving scenes for autonomous driving applications, though it is incremental as it builds on existing reconstruction methods.
The paper tackles the problem of limited realism in altered or novel driving scenarios from existing 3D scene reconstructions by introducing MADrive, a memory-augmented framework that replaces observed vehicles with similar 3D assets from a dataset of ~70K car videos, enabling photorealistic synthesis of substantially changed configurations.
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/