Long-term Temporal Convolutions for Action Recognition - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles IEEE Transactions on Pattern Analysis and Machine Intelligence Year : 2018

Long-term Temporal Convolutions for Action Recognition

Abstract

Typical human actions last several seconds and exhibit characteristic spatio-temporal structure. Recent methods attempt to capture this structure and learn action representations with convolutional neural networks. Such representations, however, are typically learned at the level of a few video frames failing to model actions at their full temporal extent. In this work we learn video representations using neural networks with long-term temporal convolutions (LTC). We demonstrate that LTC-CNN models with increased temporal extents improve the accuracy of action recognition. We also study the impact of different low-level representations, such as raw values of video pixels and optical flow vector fields and demonstrate the importance of high-quality optical flow estimation for learning accurate action models. We report state-of-the-art results on two challenging benchmarks for human action recognition UCF101 (92.7%) and HMDB51 (67.2%).
Fichier principal
Vignette du fichier
VarolPAMI2017.pdf (2.33 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01241518 , version 2 (15-04-2016)
hal-01241518 , version 3 (02-06-2017)

Identifiers

Cite

Gül Varol, Ivan Laptev, Cordelia Schmid. Long-term Temporal Convolutions for Action Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40 (6), pp.1510-1517. ⟨10.1109/TPAMI.2017.2712608⟩. ⟨hal-01241518v3⟩
8134 View
2175 Download

Altmetric

Share

Gmail Facebook X LinkedIn More