TG-Pose: Delving into Topology and Geometry
for Category-level Object Pose Estimation
TG-Pose: Delving into Topology and Geometry
for Category-level Object Pose Estimation
Yue Zhan, Xin Wang, Lang Nie, Yang Zhao, Tangwen Yang, and Qiuqi Ruan
Category-level 6D object pose estimation aims to estimate the pose and size of unseen objects with known categories. Existing methods mainly focus on capturing geometric features to handle shape variations, and are prone to failure in occlusion and noisy environments. In this paper, we propose TG-Pose, a unified pose estimation framework that delves into topology and geometry to deal with the above issues. To exploit topological properties, we first propose a topological feature predictor and a topological label generator to dig into the underlying structural details from the encoded features using persistent homology. Then, the topological and geometric features are employed to facilitate the symmetry reconstruction of the original point cloud to obtain reliable and coherent object shape, which, in turn, guide the pose estimation. For each object category, we construct geometric and topological templates by leveraging inherent intra-class similarities. These templates enhance the reliability of pose estimation and the completeness of object structure through geometric alignment and topological guidance, especially when handling incomplete objects. Moreover, a pose-aware enhancement strategy is designed to enhance the encoder in learning pose sensitive features and robustness to noisy point clouds. Experimental results show that TG-Pose outperforms the state-of-the-art solutions on public benchmarks and achieves better generalization in real-world dataset.
Overall architecture of the proposed TG-COPE.
Qualitative results of our method, GPV-Pose and HS-Pose on REAL275 dataset.
Qualitative results of our method and HS-Pose on CAMERA275 dataset.
Visualization results of symmetry reconstruction and topology prediction on REAL275 dataset.
bottle
bowl
camera
can
laptop
mug
Dynamic visualization of symmetry reconstruction results on REAL275 dataset, where bottle, bowl, can, and mug without handle conform to Rotational Symmetry, mug with handle and laptop conform to Reflection Symmetry. Blue represents the input point cloud, and red represents the reconstruction of the point cloud.
Contact
If you have any questions, please feel free to contact Yue Zhan.