no code implementations • 9 Mar 2023 • Rasmus Laurvig Haugaard, Frederik Hagelskjær, Thorbjørn Mosekjær Iversen
Pose estimation is usually approached by seeking the single best estimate of an object's pose, but this approach is ill-suited for tasks involving visual ambiguity.
no code implementations • 3 Oct 2022 • Rasmus Laurvig Haugaard, Thorbjørn Mosekjær Iversen
We present a multi-view pose estimation method which aggregates learned 2D-3D distributions from multiple views for both the initial estimate and optional refinement.
no code implementations • 20 Sep 2022 • Thorbjørn Mosekjær Iversen, Rasmus Laurvig Haugaard, Anders Glent Buch
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.
no code implementations • CVPR 2022 • Rasmus Laurvig Haugaard, Anders Glent Buch
We present an approach to learn dense, continuous 2D-3D correspondence distributions over the surface of objects from data with no prior knowledge of visual ambiguities like symmetry.