Making a Case for Learning Motion Representations with Phase

None, None; None, None

Making a Case for Learning Motion Representations with Phase

Conference Paper (2016)

Author(s)

Silvia Pintea (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Jan van Gemert (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Research Group

Pattern Recognition and Bioinformatics

DOI related publication

https://doi.org/10.1007/978-3-319-49409-8_8 Final published version

To reference this document use

https://resolver.tudelft.nl/uuid:5f86396e-5d75-4661-af58-d0d4a08739cf

More Info

expand_more

Publication Year

2016

Language

English

Research Group

Pattern Recognition and Bioinformatics

Pages (from-to)

55-64

Publisher

Springer

ISBN (print)

978-3-319-49408-1

ISBN (electronic)

978-3-319-49409-8

Event

ECCV 2016 (2016-10-08 - 2016-10-16), Amsterdam, Netherlands

Downloads counter

117

Abstract

This work advocates Eulerian motion representation learning over the current standard Lagrangian optical flow model. Eulerian motion is well captured by using phase, as obtained by decomposing the image through a complex-steerable pyramid. We discuss the gain of Eulerian motion in a set of practical use cases: (i) action recognition, (ii) motion prediction in static images, (iii) motion transfer in static images and, (iv) motion transfer in video. For each task we motivate the phase-based direction and provide a possible approach.