Omni6D: Large-Vocabulary 3D Object Dataset for Category-Level 6D Object Pose Estimation

2024-09-26

European Conference on Computer Vision

AI Summary

Existing 6D pose datasets suffer from narrow category coverage (<20 classes), limited sample size, and insufficient real-world challenges (e.g., occlusion), severely hindering category-level generalization. To address this, we introduce the first large-scale, wide-coverage RGB-D 6D pose dataset—comprising 166 object categories and over 800,000 samples—and propose three key innovations: (1) a symmetry-aware pose error metric for robust evaluation of symmetric objects; (2) a unified modeling framework based on canonical-pose normalization; and (3) a cross-dataset knowledge transfer fine-tuning strategy tailored for large-vocabulary settings. Experiments demonstrate substantial improvements in category-level generalization under occlusion and complex backgrounds. Our dataset and methodology establish a unified benchmark for over 40 state-of-the-art algorithms, advancing both academic research and industrial deployment of general-purpose 6D pose estimation.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

6D object pose estimation aims at determining an object's translation, rotation, and scale, typically from a single RGBD image. Recent advancements have expanded this estimation from instance-level to category-level, allowing models to generalize across unseen instances within the same category. However, this generalization is limited by the narrow range of categories covered by existing datasets, such as NOCS, which also tend to overlook common real-world challenges like occlusion. To tackle these challenges, we introduce Omni6D, a comprehensive RGBD dataset featuring a wide range of categories and varied backgrounds, elevating the task to a more realistic context. 1) The dataset comprises an extensive spectrum of 166 categories, 4688 instances adjusted to the canonical pose, and over 0.8 million captures, significantly broadening the scope for evaluation. 2) We introduce a symmetry-aware metric and conduct systematic benchmarks of existing algorithms on Omni6D, offering a thorough exploration of new challenges and insights. 3) Additionally, we propose an effective fine-tuning approach that adapts models from previous datasets to our extensive vocabulary setting. We believe this initiative will pave the way for new insights and substantial progress in both the industrial and academic fields, pushing forward the boundaries of general 6D pose estimation.

Problem

• Expands 6D pose estimation to 166 diverse object categories
• Addresses occlusion and real-world challenges in pose estimation
• Introduces symmetry-aware metrics for improved algorithm evaluation

Innovation

• Introduces Omni6D dataset with 166 diverse categories
• Proposes symmetry-aware metric for pose evaluation
• Develops fine-tuning approach for large-vocabulary adaptation