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

We present a new multi-stream 3D mesh reconstruction network (MSMR-Net) forhand pose estimation from a single RGB image. Our model consists of an imageencoder followed by a mesh-convolution decoder composed of connected graphconvolution layers. In contrast to previous models that form a single meshdecoding path, our decoder network incorporates multiple cross-resolutiontrajectories that are executed in parallel. Thus, global and local informationare shared to form rich decoding representations at minor additional parametercost compared to the single trajectory network. We demonstrate theeffectiveness of our method in hand-hand and hand-object interaction scenariosat various levels of interaction. To evaluate the former scenario, we propose amethod to generate RGB images of closely interacting hands. Moreoever, wesuggest a metric to quantify the degree of interaction and show that close handinteractions are particularly challenging. Experimental results show that theMSMR-Net outperforms existing algorithms on the hand-object FreiHAND dataset aswell as on our own hand-hand dataset.

Parallel mesh reconstruction streams for pose estimation of interacting hands