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 CNN shows its strong ability in learning spatiotemporal representation inrecent video recognition tasks. However, inflating 2D convolution to 3Dinevitably introduces additional computational costs, making it cumbersome inpractical deployment. We consider whether there is a way to equip theconventional 2D convolution with temporal vision no requiring expanding itskernel. To this end, we propose the video shuffle, a parameter-free plug-incomponent that efficiently reallocates the inputs of 2D convolution so that itsreceptive field can be extended to the temporal dimension. In practical, videoshuffle firstly divides each frame feature into multiple groups and thenaggregate the grouped features via temporal shuffle operation. This allows thefollowing 2D convolution aggregate the global spatiotemporal features. Theproposed video shuffle can be flexibly inserted into popular 2D CNNs, formingthe Video Shuffle Networks (VSN). With a simple yet efficient implementation,VSN performs surprisingly well on temporal modeling benchmarks. In experiments,VSN not only gains non-trivial improvements on Kinetics and Moments in Time,but also achieves state-of-the-art performance on Something-Something-V1,Something-Something-V2 datasets.

Learning Efficient Video Representation with Video Shuffle Networks