Research Analyzer
← Back ICRA 2026

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

PDF

AI summary

Key figure (auto-extracted from paper)
The newly introduced IndustryShapes dataset reveals that current state-of-the-art 6D pose estimation methods still struggle significantly with the challenging properties and realistic conditions of industrial assembly environments.
6D pose estimation RGB-D dataset industrial robotics novel object detection benchmark dataset robotic perception

Problem

Existing 6D pose estimation datasets primarily focus on household objects, controlled laboratory settings, or bin-picking scenarios, leaving a gap for realistic industrial assembly benchmarks that support both instance-level and novel-object pose estimation.

Approach

The authors introduce IndustryShapes, an RGB-D dataset featuring five industrially relevant objects with challenging properties like reflectivity and symmetry, captured in realistic assembly scenes. The dataset is split into a classic set for instance-level methods and an extended set with static onboarding sequences for novel-object approaches, alongside comprehensive benchmarks of state-of-the-art algorithms.

Key results

  • Introduction of a new RGB-D dataset with five challenging industrial objects in realistic assembly scenes
  • Provision of two complementary sets: a classic set for instance-level methods and an extended set with static onboarding sequences for novel-object pose estimation
  • Comprehensive benchmarking of state-of-the-art instance-level and novel-object pose estimation methods
  • Demonstration of significant performance gaps in current algorithms, highlighting the need for more robust industrial perception models

Why it matters

It bridges the gap between controlled lab research and real-world manufacturing deployment, providing a critical testbed for robotic perception researchers and industrial automation developers.

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 re- search 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 challeng- ing 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 es- timation, 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. This work was funded by the HEU programme SOPRANO (GA No 101120990) and PANDORA (GA No 101135775). Special thanks to Stel- lantis—Centro Ricerche FIAT (CRF) for supporting the dataset collection. All authors are with the National Technical University of Athens (NTUA), Greece (e-mail:{psapoutzoglou, orestisvaggelis, azacharia, vsartinas, mpat- eraki}@mail.ntua.gr)

Index terms

Performance Evaluation and Benchmarking Industrial Robots

Related papers