MIM4D: Masked Modeling with Multi-View Video for Autonomous Driving Representation Learning
This work addresses the problem of scalable representation learning for autonomous driving systems, offering a novel pre-training method that is incremental in improving existing tasks.
The paper tackles the challenge of learning robust visual representations from multi-view video data for autonomous driving by proposing MIM4D, a pre-training paradigm using dual masked image modeling that leverages spatial and temporal relations, achieving state-of-the-art performance on the nuScenes dataset with improvements such as 8.7% IoU on BEV segmentation and 3.5% mAP on 3D object detection.
Learning robust and scalable visual representations from massive multi-view video data remains a challenge in computer vision and autonomous driving. Existing pre-training methods either rely on expensive supervised learning with 3D annotations, limiting the scalability, or focus on single-frame or monocular inputs, neglecting the temporal information. We propose MIM4D, a novel pre-training paradigm based on dual masked image modeling (MIM). MIM4D leverages both spatial and temporal relations by training on masked multi-view video inputs. It constructs pseudo-3D features using continuous scene flow and projects them onto 2D plane for supervision. To address the lack of dense 3D supervision, MIM4D reconstruct pixels by employing 3D volumetric differentiable rendering to learn geometric representations. We demonstrate that MIM4D achieves state-of-the-art performance on the nuScenes dataset for visual representation learning in autonomous driving. It significantly improves existing methods on multiple downstream tasks, including BEV segmentation (8.7% IoU), 3D object detection (3.5% mAP), and HD map construction (1.4% mAP). Our work offers a new choice for learning representation at scale in autonomous driving. Code and models are released at https://github.com/hustvl/MIM4D