The goal of this project is to jointly refine initial hand and object pose estimates given a single RGB image, including multiple goals: (i) evaluation of differentiable rendering as a feedback mechanism for iterative refinement, (ii) evaluation of existing network training objectives with proposal of new ones, (iii) a complete method proposal and (iv) implementation of the propsed method and an evaluation against the state-of-the-art.
|Effective start/end date||1/09/21 → 28/02/22|
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.