Learnable Triangulation of Human Pose

July 18, 2019 11:00 am - 12:00 pm

Bookmark and Share

Yuri Malkov

We present two novel solutions for multi-view 3D human pose estimation based on new learnable triangulation methods that combine 3D information from multiple 2D views. Crucially, both approaches are end-to-end differentiable, which allows us to directly optimize the target metric. We demonstrate transferability of the solutions across datasets and significantly improve the multi-view state of the art on the Human3.6M dataset.