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

Recent studies on the automatic detection of facial action unit (AU) haveextensively relied on large-sized annotations. However, manually AU labeling isdifficult, time-consuming, and costly. Most existing semi-supervised worksignore the informative cues from the temporal domain, and are highly dependenton densely annotated videos, making the learning process less efficient. Toalleviate these problems, we propose a deep semi-supervised frameworkKnowledge-Spreader (KS), which differs from conventional methods in twoaspects. First, rather than only encoding human knowledge as constraints, KSalso learns the Spatial-Temporal AU correlation knowledge in order tostrengthen its out-of-distribution generalization ability. Second, we approachKS by applying consistency regularization and pseudo-labeling in multiplestudent networks alternately and dynamically. It spreads the spatial knowledgefrom labeled frames to unlabeled data, and completes the temporal informationof partially labeled video clips. Thus, the design allows KS to learn AUdynamics from video clips with only one label allocated, which significantlyreduce the requirements of using annotations. Extensive experiments demonstratethat the proposed KS achieves competitive performance as compared to the stateof the arts under the circumstances of using only 2% labels on BP4D and 5%labels on DISFA. In addition, we test it on our newly developed large-scalecomprehensive emotion database, which contains considerable samples acrosswell-synchronized and aligned sensor modalities for easing the scarcity issueof annotations and identities in human affective computing. The new databasewill be released to the research community.

Knowledge-Spreader: Learning Facial Action Unit Dynamics with Extremely Limited Labels