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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Existing RNN-based approaches for action recognition from depth sequencesrequire either skeleton joints or hand-crafted depth features as inputs. Anend-to-end manner, mapping from raw depth maps to action classes, isnon-trivial to design due to the fact that: 1) single channel map lacks texturethus weakens the discriminative power; 2) relatively small set of depthtraining data. To address these challenges, we propose to learn an RNN drivenby privileged information (PI) in three-steps: An encoder is pre-trained tolearn a joint embedding of depth appearance and PI (i.e. skeleton joints). Thelearned embedding layers are then tuned in the learning step, aiming tooptimize the network by exploiting PI in a form of multi-task loss. However,exploiting PI as a secondary task provides little help to improve theperformance of a primary task (i.e. classification) due to the gap betweenthem. Finally, a bridging matrix is defined to connect two tasks by discoveringlatent PI in the refining step. Our PI-based classification loss maintains aconsistency between latent PI and predicted distribution. The latent PI andnetwork are iteratively estimated and updated in an expectation-maximizationprocedure. The proposed learning process provides greater discriminative powerto model subtle depth difference, while helping avoid overfitting the scarcertraining data. Our experiments show significant performance gains overstate-of-the-art methods on three public benchmark datasets and our newlycollected Blanket dataset.

Learning and Refining of Privileged Information-based RNNs for Action Recognition from Depth Sequences