Object Pose

Object pose estimation is a key problem for Augmented Reality as well as Robotic applications.

In our CVPR 2022 paper, “ZebraPose: Coarse to Fine Surface Encoding for 6DoF Object Pose Estimation we proposed a formulation of the 6DoF problem as a hierarchical binary surface encoding prediction.

In the follow-up work HiPose at CVPR 2024, we extended ZebraPose to directly make use of RGB-D data.

The article Resolving Symmetry Ambiguity in Correspondence-Based Methods for Instance-Level Object Pose Estimation (IEEE Transactions on Image Processing, 2025) adapts the surface encodings of ZebraPose for symmetric objects.

Previously, we applied our object pose estimation methods to Augmented Reality assembly applications where we estimate the current pose and state of an assembly process ( ISMAR 2019 paper )