RFFCE: Residual Feature Fusion and Confidence Evaluation Network for 6DoF Pose Estimation
Qiwei Meng, Shanshan Ji, Shiqiang Zhu, Tianlei Jin, Te Li, Jason Gu, Wei Song
Abstract
In this paper, we propose a novel RGBD-based object 6DoF pose estimation network - RFFCE. It is a two-stage method that firstly leverages deep neural networks for feature extraction and object points matching, and then the geometric principles are utilized for final pose computation. Our approach consists of three primary innovations: residual feature fusion for representative RGBD feature extraction; confidence evaluation and confidence-based paired points offsets regression for self-evaluation and self-optimization respectively. Their effectiveness is verified through an ablation study, and our RFFCE achieves the SOTA performance on LineMOD, Occlusion-LineMOD and YCB-Video datasets. Additionally, we also conduct a real-world object grasping experiment for visualization and qualitative evaluation of the RFFCE.
BibTeX
@inproceedings{icra2023_rffceresidualfea,
title = {RFFCE: Residual Feature Fusion and Confidence Evaluation Network for 6DoF Pose Estimation},
author = {Qiwei Meng and Shanshan Ji and Shiqiang Zhu and Tianlei Jin and Te Li and Jason Gu and Wei Song},
booktitle = {ICRA 2023},
year = {2023}
}