Lifting Unlabeled Internet-level Data for 3D Scene Understanding
This addresses the problem of expensive data annotation for 3D scene understanding, offering an incremental approach to leverage abundant web data.
The paper tackles the scarcity of annotated 3D scene data by using unlabeled internet videos to automatically generate training data, enabling models to achieve strong zero-shot performance and further improvements after fine-tuning on tasks like 3D object detection and visual question answering.
Annotated 3D scene data is scarce and expensive to acquire, while abundant unlabeled videos are readily available on the internet. In this paper, we demonstrate that carefully designed data engines can leverage web-curated, unlabeled videos to automatically generate training data, to facilitate end-to-end models in 3D scene understanding alongside human-annotated datasets. We identify and analyze bottlenecks in automated data generation, revealing critical factors that determine the efficiency and effectiveness of learning from unlabeled data. To validate our approach across different perception granularities, we evaluate on three tasks spanning low-level perception, i.e., 3D object detection and instance segmentation, to high-evel reasoning, i.e., 3D spatial Visual Question Answering (VQA) and Vision-Lanugage Navigation (VLN). Models trained on our generated data demonstrate strong zero-shot performance and show further improvement after finetuning. This demonstrates the viability of leveraging readily available web data as a path toward more capable scene understanding systems.