Extracting Action Sensitive Features to Facilitate Weakly-Supervised Action Localization - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2019

Extracting Action Sensitive Features to Facilitate Weakly-Supervised Action Localization

Le Wang
  • Fonction : Auteur
  • PersonId : 1033453
Qilin Zhang
  • Fonction : Auteur

Résumé

Weakly-supervised temporal action localization has attracted much attention among researchers in video content analytics, thanks to its relaxed requirements of video-level annotations instead of frame-level labels. However, many current weakly-supervised action localization methods depend heavily on naive feature combination and empirical thresholds to determine temporal action boundaries, which is practically feasible but could still be sub-optimal. Inspired by the momentum term, we propose a general-purpose action recognition criterion that replaces explicit empirical thresholds. Based on such criterion, we analyze different combination of streams and propose the Action Sensitive Extractor (ASE) that produces action sensitive features. Our ASE sets temporal stream as main stream and extends with complementary spatial streams. We build our Action Sensitive Network (ASN) and evaluate on THUMOS14 and ActivityNet1.2 with different selection method. Our network yields state-of-art performance in both datasets.
Fichier principal
Vignette du fichier
483292_1_En_15_Chapter.pdf (385.23 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02331300 , version 1 (24-10-2019)

Licence

Paternité

Identifiants

Citer

Zijian Kang, Le Wang, Ziyi Liu, Qilin Zhang, Nanning Zheng. Extracting Action Sensitive Features to Facilitate Weakly-Supervised Action Localization. 15th IFIP International Conference on Artificial Intelligence Applications and Innovations (AIAI), May 2019, Hersonissos, Greece. pp.188-201, ⟨10.1007/978-3-030-19823-7_15⟩. ⟨hal-02331300⟩
45 Consultations
24 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More