Action localization with tubelets from motion

Mihir Jain 1 Jan van Gemert 2 Hervé Jégou 1 Patrick Bouthemy 3 Cees Snoek 2
1 TEXMEX - Multimedia content-based indexing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : This paper considers the problem of action localization, where the objective is to determine when and where certain actions appear. We introduce a sampling strategy to produce 2D+t sequences of bounding boxes, called tubelets. Compared to state-of-the-art alternatives, this drastically reduces the number of hypotheses that are likely to include the action of interest. Our method is inspired by a recent technique introduced in the context of image localization. Beyond considering this technique for the first time for videos, we revisit this strategy for 2D+t sequences obtained from super-voxels. Our sampling strategy advantageously exploits a criterion that reflects how action related motion deviates from background motion. We demonstrate the interest of our approach by extensive experiments on two public datasets: UCF Sports and MSR-II. Our approach significantly outperforms the state-of-the-art on both datasets, while restricting the search of actions to a fraction of possible bounding box sequences.
Document type :
Conference papers
Complete list of metadatas

Cited literature [35 references]  Display  Hide  Download

https://hal.inria.fr/hal-00996844
Contributor : Hervé Jégou <>
Submitted on : Tuesday, May 27, 2014 - 9:31:58 AM
Last modification on : Friday, November 16, 2018 - 1:24:43 AM
Long-term archiving on : Wednesday, August 27, 2014 - 11:01:12 AM

File

jain_cvpr2014_tubelets.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00996844, version 1

Citation

Mihir Jain, Jan van Gemert, Hervé Jégou, Patrick Bouthemy, Cees Snoek. Action localization with tubelets from motion. CVPR - International Conference on Computer Vision and Pattern Recognition, Jun 2014, Columbus, United States. ⟨hal-00996844⟩

Share

Metrics

Record views

1377

Files downloads

525