English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

3D reconstruction of hand-object manipulations is important for emulatinghuman actions. Most methods dealing with challenging object manipulationscenarios, focus on hands reconstruction in isolation, ignoring physical andkinematic constraints due to object contact. Some approaches produce morerealistic results by jointly reconstructing 3D hand-object interactions.However, they focus on coarse pose estimation or rely upon known hand andobject shapes. We propose the first approach for realistic 3D hand-object shapeand pose reconstruction from a single depth map. Unlike previous work, ourvoxel-based reconstruction network regresses the vertex coordinates of a handand an object and reconstructs more realistic interaction. Our pipelineadditionally predicts voxelized hand-object shapes, having a one-to-one mappingto the input voxelized depth. Thereafter, we exploit the graph nature of thehand and object shapes, by utilizing the recent GraFormer network withpositional embedding to reconstruct shapes from template meshes. In addition,we show the impact of adding another GraFormer component that refines thereconstructed shapes based on the hand-object interactions and its ability toreconstruct more accurate object shapes. We perform an extensive evaluation onthe HO-3D and DexYCB datasets and show that our method outperforms existingapproaches in hand reconstruction and produces plausible reconstructions forthe objects

ShapeGraFormer: GraFormer-Based Network for Hand-Object Reconstruction from a Single Depth Map