Hierarchical Attention and Graph Neural Networks: Toward Drift-Free Pose Estimation

by   Kathia Melbouci, et al.

The most commonly used method for addressing 3D geometric registration is the iterative closet-point algorithm, this approach is incremental and prone to drift over multiple consecutive frames. The Common strategy to address the drift is the pose graph optimization subsequent to frame-to-frame registration, incorporating a loop closure process that identifies previously visited places. In this paper, we explore a framework that replaces traditional geometric registration and pose graph optimization with a learned model utilizing hierarchical attention mechanisms and graph neural networks. We propose a strategy to condense the data flow, preserving essential information required for the precise estimation of rigid poses. Our results, derived from tests on the KITTI Odometry dataset, demonstrate a significant improvement in pose estimation accuracy. This improvement is especially notable in determining rotational components when compared with results obtained through conventional multi-way registration via pose graph optimization. The code will be made available upon completion of the review process.


page 1

page 2

page 3

page 4


GeoLCR: Attention-based Geometric Loop Closure and Registration

We present a novel algorithm for learning-based loop-closure for SLAM (s...

Drift Reduction for Monocular Visual Odometry of Intelligent Vehicles using Feedforward Neural Networks

In this paper, an approach for reducing the drift in monocular visual od...

Learning Dynamics via Graph Neural Networks for Human Pose Estimation and Tracking

Multi-person pose estimation and tracking serve as crucial steps for vid...

SEM-GAT: Explainable Semantic Pose Estimation using Learned Graph Attention

This paper proposes a GNN-based method for exploiting semantics and loca...

Learning Skeletal Graph Neural Networks for Hard 3D Pose Estimation

Various deep learning techniques have been proposed to solve the single-...

Trajectory Space Factorization for Deep Video-Based 3D Human Pose Estimation

Existing deep learning approaches on 3d human pose estimation for videos...

Suspicious Massive Registration Detection via Dynamic Heterogeneous Graph Neural Networks

Massive account registration has raised concerns on risk management in e...

Please sign up or login with your details

Forgot password? Click here to reset