Towards Two-view 6D Object Pose Estimation: A Comparative Study on Fusion Strategy

TL;DR

This paper investigates fusion strategies for RGB-based 6D object pose estimation, proposing a framework that learns implicit 3D information from RGB images and demonstrating that mid-fusion yields the best results, narrowing the gap with RGBD methods.

Contribution

It introduces a novel framework that learns implicit 3D features from RGB images and compares three fusion strategies, identifying mid-fusion as the most effective for pose estimation.

Findings

Mid-fusion approach best restores 3D keypoints.

Proposed method outperforms existing RGB-based methods.

Achieves comparable results with RGBD-based methods.

Abstract

Current RGB-based 6D object pose estimation methods have achieved noticeable performance on datasets and real world applications. However, predicting 6D pose from single 2D image features is susceptible to disturbance from changing of environment and textureless or resemblant object surfaces. Hence, RGB-based methods generally achieve less competitive results than RGBD-based methods, which deploy both image features and 3D structure features. To narrow down this performance gap, this paper proposes a framework for 6D object pose estimation that learns implicit 3D information from 2 RGB images. Combining the learned 3D information and 2D image features, we establish more stable correspondence between the scene and the object models. To seek for the methods best utilizing 3D information from RGB inputs, we conduct an investigation on three different approaches, including Early- Fusion, Mid-Fusion, and Late-Fusion. We ascertain the Mid- Fusion approach is the best approach to restore the most precise 3D keypoints useful for object pose estimation. The experiments show that our method outperforms state-of-the-art RGB-based methods, and achieves comparable results with RGBD-based methods.