Lifting Unlabeled Internet-level Data for 3D Scene Understanding

TL;DR

This paper presents a method to automatically generate training data from unlabeled web videos to improve 3D scene understanding models across various perception tasks.

Contribution

It introduces a data engine that leverages web videos for training, addressing data scarcity and demonstrating effectiveness across multiple perception tasks.

Findings

Models trained on generated data show strong zero-shot performance.

Fine-tuning further improves model accuracy.

The approach reveals key bottlenecks in automated data generation.

Abstract

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.