SEM-GAT: Explainable Semantic Pose Estimation using Learned Graph Attention
This work addresses pose estimation for autonomous vehicles, but it is incremental as it builds on existing GNN methods with a novel graph structure.
The paper tackles the problem of lidar-based pose estimation by using a Graph Neural Network to exploit semantic and geometric features for pointcloud registration, achieving competitive accuracy and higher track smoothness on the KITTI odometry dataset with fewer parameters.
This paper proposes a Graph Neural Network(GNN)-based method for exploiting semantics and local geometry to guide the identification of reliable pointcloud registration candidates. Semantic and morphological features of the environment serve as key reference points for registration, enabling accurate lidar-based pose estimation. Our novel lightweight static graph structure informs our attention-based node aggregation network by identifying semantic-instance relationships, acting as an inductive bias to significantly reduce the computational burden of pointcloud registration. By connecting candidate nodes and exploiting cross-graph attention, we identify confidence scores for all potential registration correspondences and estimate the displacement between pointcloud scans. Our pipeline enables introspective analysis of the model's performance by correlating it with the individual contributions of local structures in the environment, providing valuable insights into the system's behaviour. We test our method on the KITTI odometry dataset, achieving competitive accuracy compared to benchmark methods and a higher track smoothness while relying on significantly fewer network parameters.