Ki-Pode: Keypoint-based Implicit Pose Distribution Estimation of Rigid Objects
This addresses the need for uncertainty-aware pose estimation in computer vision, particularly for applications where failure costs are high, though it appears incremental by extending existing keypoint-based methods to distribution estimation.
The authors tackled the problem of 6D pose estimation for rigid objects by proposing a method to estimate full pose distributions instead of single best estimates, which reliably handles visual ambiguities due to symmetries or occlusions, as demonstrated on YCB-V and T-LESS datasets.
The estimation of 6D poses of rigid objects is a fundamental problem in computer vision. Traditionally pose estimation is concerned with the determination of a single best estimate. However, a single estimate is unable to express visual ambiguity, which in many cases is unavoidable due to object symmetries or occlusion of identifying features. Inability to account for ambiguities in pose can lead to failure in subsequent methods, which is unacceptable when the cost of failure is high. Estimates of full pose distributions are, contrary to single estimates, well suited for expressing uncertainty on pose. Motivated by this, we propose a novel pose distribution estimation method. An implicit formulation of the probability distribution over object pose is derived from an intermediary representation of an object as a set of keypoints. This ensures that the pose distribution estimates have a high level of interpretability. Furthermore, our method is based on conservative approximations, which leads to reliable estimates. The method has been evaluated on the task of rotation distribution estimation on the YCB-V and T-LESS datasets and performs reliably on all objects.