Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

IndustryShapes: An RGB-D Benchmark dataset for 6D object pose estimation of industrial assembly components and tools

Panagiotis Sapoutzoglou, Orestis Vaggelis, Athina Zacharia, Evangelos Sartinas, Maria Pateraki

TL;DR

IndustryShapes tackles the challenge of real-world industrial 6D pose estimation by providing a realistic RGB-D benchmark with five challenging industrial objects, organized into Classic (instance-level) and Extended (novel-object) sets, plus RGB-D onboarding sequences. The dataset supports model-based, model-free, and sequence-driven approaches and uses the BOP protocol to evaluate pose, detection, and segmentation with standardized metrics (VSD, MSSD, MSPD, ADD, AR, mAP). Baselines from EPOS, ZebraPose, DOPE, FoundPose, FoundationPose, CNOS, and SAM-6D reveal substantial domain gaps and the need for progress to handle industrial realism, occlusion, and reflective geometries. Overall, IndustryShapes provides a rigorous, deployment-relevant testbed that highlights current limitations and stimulates the development of robust industrial 6D pose estimation methods.

Abstract

We introduce IndustryShapes, a new RGB-D benchmark dataset of industrial tools and components, designed for both instance-level and novel object 6D pose estimation approaches. The dataset provides a realistic and application-relevant testbed for benchmarking these methods in the context of industrial robotics bridging the gap between lab-based research and deployment in real-world manufacturing scenarios. Unlike many previous datasets that focus on household or consumer products or use synthetic, clean tabletop datasets, or objects captured solely in controlled lab environments, IndustryShapes introduces five new object types with challenging properties, also captured in realistic industrial assembly settings. The dataset has diverse complexity, from simple to more challenging scenes, with single and multiple objects, including scenes with multiple instances of the same object and it is organized in two parts: the classic set and the extended set. The classic set includes a total of 4,6k images and 6k annotated poses. The extended set introduces additional data modalities to support the evaluation of model-free and sequence-based approaches. To the best of our knowledge, IndustryShapes is the first dataset to offer RGB-D static onboarding sequences. We further evaluate the dataset on a representative set of state-of-the art methods for instance-based and novel object 6D pose estimation, including also object detection, segmentation, showing that there is room for improvement in this domain. The dataset page can be found in https://pose-lab.github.io/IndustryShapes.

IndustryShapes: An RGB-D Benchmark dataset for 6D object pose estimation of industrial assembly components and tools

TL;DR

IndustryShapes tackles the challenge of real-world industrial 6D pose estimation by providing a realistic RGB-D benchmark with five challenging industrial objects, organized into Classic (instance-level) and Extended (novel-object) sets, plus RGB-D onboarding sequences. The dataset supports model-based, model-free, and sequence-driven approaches and uses the BOP protocol to evaluate pose, detection, and segmentation with standardized metrics (VSD, MSSD, MSPD, ADD, AR, mAP). Baselines from EPOS, ZebraPose, DOPE, FoundPose, FoundationPose, CNOS, and SAM-6D reveal substantial domain gaps and the need for progress to handle industrial realism, occlusion, and reflective geometries. Overall, IndustryShapes provides a rigorous, deployment-relevant testbed that highlights current limitations and stimulates the development of robust industrial 6D pose estimation methods.

Abstract

We introduce IndustryShapes, a new RGB-D benchmark dataset of industrial tools and components, designed for both instance-level and novel object 6D pose estimation approaches. The dataset provides a realistic and application-relevant testbed for benchmarking these methods in the context of industrial robotics bridging the gap between lab-based research and deployment in real-world manufacturing scenarios. Unlike many previous datasets that focus on household or consumer products or use synthetic, clean tabletop datasets, or objects captured solely in controlled lab environments, IndustryShapes introduces five new object types with challenging properties, also captured in realistic industrial assembly settings. The dataset has diverse complexity, from simple to more challenging scenes, with single and multiple objects, including scenes with multiple instances of the same object and it is organized in two parts: the classic set and the extended set. The classic set includes a total of 4,6k images and 6k annotated poses. The extended set introduces additional data modalities to support the evaluation of model-free and sequence-based approaches. To the best of our knowledge, IndustryShapes is the first dataset to offer RGB-D static onboarding sequences. We further evaluate the dataset on a representative set of state-of-the art methods for instance-based and novel object 6D pose estimation, including also object detection, segmentation, showing that there is room for improvement in this domain. The dataset page can be found in https://pose-lab.github.io/IndustryShapes.
Paper Structure (15 sections, 3 figures, 7 tables)

This paper contains 15 sections, 3 figures, 7 tables.

Table of Contents

  1. INTRODUCTION
  2. Related work
  3. Datasets
  4. Methods
  5. The IndustryShapes dataset
  6. Objects and 3D models
  7. Dataset structure and Composition
  8. 6D pose annotation of real image sequences
  9. Benchmark
  10. Evaluation metrics
  11. Baselines and evaluation setup
  12. Performance evaluation
  13. 6D Pose Estimation
  14. Detection and segmentation
  15. Conclusions

Figures (3)

  • Figure 1: Overview of the IndustryShapes dataset. Five industrial tools/components with challenging properties (weak/absent texture, symmetries, thin and reflective parts) are captured in realistic industrial environment scenes. Data are organized into two complementary sets: Classic set for instance-level evaluation and Extended set tailored for novel-object methods, including RGB-D static onboarding sequences.
  • Figure 2: Pose distribution per Object. Visualization of the overall spherical viewpoint coverage of the complete IndustryShapes dataset in Mollweide projection, indicating the density and pose variation.
  • Figure 3: Distribution of object-to-camera distances for annotated poses, grouped by object (1 to 5 from left to right). Top row: annotated poses in the training (blue) and test (magenta) data of the classic set. Bottom row: annotated poses of the classic set (orange) and the extended set (yellow).