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 propose a self-supervised learning method to jointly reason about spatialand temporal context for video recognition. Recent self-supervised approacheshave used spatial context [9, 34] as well as temporal coherency [32] but acombination of the two requires extensive preprocessing such as trackingobjects through millions of video frames [59] or computing optical flow todetermine frame regions with high motion [30]. We propose to combine spatialand temporal context in one self-supervised framework without any heavypreprocessing. We divide multiple video frames into grids of patches and traina network to solve jigsaw puzzles on these patches from multiple frames. So thenetwork is trained to correctly identify the position of a patch within a videoframe as well as the position of a patch over time. We also propose a novelpermutation strategy that outperforms random permutations while significantlyreducing computational and memory constraints. We use our trained network fortransfer learning tasks such as video activity recognition and demonstrate thestrength of our approach on two benchmark video action recognition datasetswithout using a single frame from these datasets for unsupervised pretrainingof our proposed video jigsaw network.

Video Jigsaw: Unsupervised Learning of Spatiotemporal Context for Video Action Recognition