6DoF Object Pose Estimation via Differentiable Proxy Voting Loss

TL;DR

This paper introduces a differentiable proxy voting loss for 6DOF object pose estimation that accounts for pixel-to-keypoint distances, improving accuracy and training efficiency.

Contribution

It proposes a novel differentiable loss function that enhances vector-field based keypoint voting by incorporating pixel-keypoint distances, enabling end-to-end training.

Findings

Significant improvement in pose estimation accuracy on LINEMOD datasets.

Faster training convergence compared to previous methods.

Effective handling of occlusions and textureless objects.

Abstract

Estimating a 6DOF object pose from a single image is very challenging due to occlusions or textureless appearances. Vector-field based keypoint voting has demonstrated its effectiveness and superiority on tackling those issues. However, direct regression of vector-fields neglects that the distances between pixels and keypoints also affect the deviations of hypotheses dramatically. In other words, small errors in direction vectors may generate severely deviated hypotheses when pixels are far away from a keypoint. In this paper, we aim to reduce such errors by incorporating the distances between pixels and keypoints into our objective. To this end, we develop a simple yet effective differentiable proxy voting loss (DPVL) which mimics the hypothesis selection in the voting procedure. By exploiting our voting loss, we are able to train our network in an end-to-end manner. Experiments on widely used datasets, i.e., LINEMOD and Occlusion LINEMOD, manifest that our DPVL improves pose estimation performance significantly and speeds up the training convergence.