Temporal Localization of Actions with Actoms

By Adrien Gaidon, Zaïd Harchaoui, Cordelia Schmid. 2013

190 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

Citations in the corpus

Towards Automatic Learning of Procedures From Web Instructional Videos (2017, background): “Early works on action detection mainly use sliding windows for proposing segments (Gaidon, Harchaoui, and Schmid 2013; Oneata, Verbeek, and Schmid 2013).”
Dense-Captioning Events in Videos (2017, background): “With the introduction of large scale activity datasets [22, 19, 12, 4], it has become possible to categorize videos into a discrete set of action categories [27, 10, 9, 40, 35].”
Temporal Segment Networks: Towards Good Practices for Deep Action Recognition (2016, methodology): “Many research works have been devoted to modeling the temporal structure for action recognition [15,16,17,29,30,18].”
A New Representation of Skeleton Sequences for 3D Action Recognition (2017, background): “To recognize a video action, the temporal information of the sequence needs to be exploited to understand the dynamics of human postures [29, 9, 44, 8, 20].”
Temporal Relational Reasoning in Videos (2017, neutral): “Besides, many previous works model the temporal structures of videos for action recognition and detection using bag of words, motion atoms, or action grammar [5, 16, 26, 6, 27].”
Towards Weakly-Supervised Action Localization (2016, background): “Initial attempts for temporal and spatio-temporal action localization are based on a sliding-window scheme and focus on improving the search complexity [11,18,19,20].”
Temporal Segment Networks for Action Recognition in Videos (2017, methodology): “been proven crucial for understanding the dynamics in traditional methods [19], [20], [21], [22], [23], it has not been consid-”
Learning to Track for Spatio-Temporal Action Localization (2015, neutral): “For temporal localization of actions, most state-of-theart approaches are based on a sliding window [7, 44].”
Discriminative Hierarchical Modeling of Spatio-temporally Composable Human Activities (2014, background): “Other researchers have proposed models for decomposing actions into short temporal motion segments [14, 25], but cannot capture spatial composition of actions.”
FineGym: A Hierarchical Video Dataset for Fine-Grained Action Understanding (2020, background): “Besides action recognition, other tasks of action understanding, including action detection and localization [14, 47, 54, 21, 16, 36], action segmentation [26, 9], and action generation [28, 41], also attract many researchers.”
Temporal Action Localization in Untrimmed Videos via Multi-stage CNNs (2016, background): “[4, 5] modeled the structure of action sequence with atomic action units (actoms).”
Online Real-Time Multiple Spatiotemporal Action Localisation and Prediction (2016, methodology): “Temporal action detection is mostly tackled using expensive sliding window [20, 5, 45, 26, 50] approaches.”
Trespassing the Boundaries: Labeling Temporal Bounds for Object Interactions in Egocentric Video (2017, background): “However, both approaches [9, 34] rely on manual annotations of actoms [9] or action segments [34], which are potentially as subjective as the temporal bounds of the actions themselves.”
Dense-Caption Matching and Frame-Selection Gating for Temporal Localization in VideoQA (2020, background): “There has been work that solely processes visual information to detect objects/actions/activity (Gaidon et al., 2013; Weinzaepfel et al., 2015; Shou et al., 2016; Dai et al., 2017; Shou et al., 2017).”
Multi-region Two-Stream R-CNN for Action Detection (2016, unknown): “For action detection, [10, 11, 29, 30] use local features to represent actions and rely on a sliding window scheme for either temporal or spatio-temporal localization.”
ActionVLAD: Learning Spatio-Temporal Aggregation for Action Classification (2017, methodology): “, their method is not end-to-end trainable and is used as a post-processing step. Other works in event detection and action classiﬁcation rely on pooling hand crafted features over segments of videos [17,30,32,37,48]. Others have also investigated pooling convolutional descriptors from video using, ”