Spatial Feature Mapping for 6DoF Object Pose Estimation

TL;DR

This paper introduces a novel approach for 6DoF object pose estimation in cluttered environments by leveraging spatial structure through graph-based modeling and spherical convolutions, improving accuracy under occlusion.

Contribution

It proposes a graph-based spatial feature mapping method combined with spherical convolution to enhance pose estimation accuracy in challenging scenarios.

Findings

Effective in cluttered backgrounds and occlusion

Improves pose accuracy on YCB-Video and LINEMOD datasets

Utilizes both depth and RGB information for robust estimation

Abstract

This work aims to estimate 6Dof (6D) object pose in background clutter. Considering the strong occlusion and background noise, we propose to utilize the spatial structure for better tackling this challenging task. Observing that the 3D mesh can be naturally abstracted by a graph, we build the graph using 3D points as vertices and mesh connections as edges. We construct the corresponding mapping from 2D image features to 3D points for filling the graph and fusion of the 2D and 3D features. Afterward, a Graph Convolutional Network (GCN) is applied to help the feature exchange among objects' points in 3D space. To address the problem of rotation symmetry ambiguity for objects, a spherical convolution is utilized and the spherical features are combined with the convolutional features that are mapped to the graph. Predefined 3D keypoints are voted and the 6DoF pose is obtained via the fitting optimization. Two scenarios of inference, one with the depth information and the other without it are discussed. Tested on the datasets of YCB-Video and LINEMOD, the experiments demonstrate the effectiveness of our proposed method.