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

Currently, the most common motion representation for action recognition isoptical flow. Optical flow is based on particle tracking which adheres to aLagrangian perspective on dynamics. In contrast to the Lagrangian perspective,the Eulerian model of dynamics does not track, but describes local changes. Forvideo, an Eulerian phase-based motion representation, using complex steerablefilters, has been successfully employed recently for motion magnification andvideo frame interpolation. Inspired by these previous works, here, we proposeslearning Eulerian motion representations in a deep architecture for actionrecognition. We learn filters in the complex domain in an end-to-end manner. Wedesign these complex filters to resemble complex Gabor filters, typicallyemployed for phase-information extraction. We propose a phase-informationextraction module, based on these complex filters, that can be used in anynetwork architecture for extracting Eulerian representations. We experimentallyanalyze the added value of Eulerian motion representations, as extracted by ourproposed phase extraction module, and compare with existing motionrepresentations based on optical flow, on the UCF101 dataset.

Using phase instead of optical flow for action recognition