Pose Estimation from RGB Images of Highly Symmetric Objects using a Novel Multi-Pose Loss and Differential Rendering

Stefan Hein Bengtson, Hampus Åström, Thomas B. Moeslund, Elin Anna Topp, Volker Krueger

Research output: Chapter in Book/Report/Conference proceedingPaper in conference proceedingpeer-review

Abstract

We propose a novel multi-pose loss function to train a neural network for 6D pose estimation, using synthetic data and evaluating it on real images. Our loss is inspired by the VSD (Visible Surface Discrepancy) metric and relies on a differentiable renderer and CAD models. This novel multi-pose approach produces multiple weighted pose estimates to avoid getting stuck in local minima. Our method resolves pose ambiguities without using predefined symmetries. It is trained only on synthetic data. We test on real-world RGB images from the T-LESS dataset, containing highly symmetric objects common in industrial settings. We show that our solution can be used to replace the codebook in a state-of-the-art approach. So far, the codebook approach has had the shortest inference time in the field. Our approach reduces inference time further while a) avoiding discretization, b) requiring a much smaller memory footprint and c) improving pose recall.
Original languageEnglish
Title of host publicationIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
PublisherIEEE - Institute of Electrical and Electronics Engineers Inc.
Edition2021
ISBN (Electronic)978-1-6654-1714-3
ISBN (Print)978-1-6654-1715-0
DOIs
Publication statusPublished - 2021 Sept
EventIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2021 - Prague, Czech Republic
Duration: 2021 Sept 272021 Oct 1

Conference

ConferenceIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2021
Country/TerritoryCzech Republic
CityPrague
Period2021/09/272021/10/01

Subject classification (UKÄ)

  • Computer Science

Free keywords

  • Training
  • Measurement
  • Solid modeling
  • image resolution
  • Pose estimation
  • neural networks
  • memory management

Fingerprint

Dive into the research topics of 'Pose Estimation from RGB Images of Highly Symmetric Objects using a Novel Multi-Pose Loss and Differential Rendering'. Together they form a unique fingerprint.

Cite this