A Hybrid Approach for 6DoF Pose Estimation

TL;DR

This paper introduces a hybrid 6DoF pose estimation method combining deep learning for object detection with geometric voting for pose recovery, optimized by automatic method selection, achieving state-of-the-art results.

Contribution

It presents a novel hybrid approach that integrates CNN-based detection with geometric pose estimation and automatic method selection for improved accuracy.

Findings

Outperforms baseline methods on BOP datasets

Achieves top speed in BOP 2020 Challenge

Significantly improves pose estimation accuracy

Abstract

We propose a method for 6DoF pose estimation of rigid objects that uses a state-of-the-art deep learning based instance detector to segment object instances in an RGB image, followed by a point-pair based voting method to recover the object's pose. We additionally use an automatic method selection that chooses the instance detector and the training set as that with the highest performance on the validation set. This hybrid approach leverages the best of learning and classic approaches, using CNNs to filter highly unstructured data and cut through the clutter, and a local geometric approach with proven convergence for robust pose estimation. The method is evaluated on the BOP core datasets where it significantly exceeds the baseline method and is the best fast method in the BOP 2020 Challenge.