Action anticipation with RBF kernelized feature mapping RNN

Yuge Shi*, Basura Fernando, Richard Hartley

*Corresponding author for this work

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    14 Citations (Scopus)

    Abstract

    We introduce a novel Recurrent Neural Network-based algorithm for future video feature generation and action anticipation called feature mapping RNN. Our novel RNN architecture builds upon three effective principles of machine learning, namely parameter sharing, Radial Basis Function kernels and adversarial training. Using only some of the earliest frames of a video, the feature mapping RNN is able to generate future features with a fraction of the parameters needed in traditional RNN. By feeding these future features into a simple multilayer perceptron facilitated with an RBF kernel layer, we are able to accurately predict the action in the video. In our experiments, we obtain 18% improvement on JHMDB-21 dataset, 6% on UCF101-24 and 13% improvement on UT-Interaction datasets over prior state-of-the-art for action anticipation.

    Original languageEnglish
    Title of host publicationComputer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings
    EditorsMartial Hebert, Vittorio Ferrari, Cristian Sminchisescu, Yair Weiss
    PublisherSpringer Verlag
    Pages305-322
    Number of pages18
    ISBN (Print)9783030012489
    DOIs
    Publication statusPublished - 2018
    Event15th European Conference on Computer Vision, ECCV 2018 - Munich, Germany
    Duration: 8 Sept 201814 Sept 2018

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume11214 LNCS
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Conference

    Conference15th European Conference on Computer Vision, ECCV 2018
    Country/TerritoryGermany
    CityMunich
    Period8/09/1814/09/18

    Fingerprint

    Dive into the research topics of 'Action anticipation with RBF kernelized feature mapping RNN'. Together they form a unique fingerprint.

    Cite this