Learning spatiotemporal graphs of human activities

By William Brendel, Sinisa Todorovic. 2011

257 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

What do we need to build explainable AI systems for the medical domain? (2017, methodology): “…techniques such as Bayesian Program Learning (Lake et al., 2015), learning models of causal relationships (Maier et al., 2010, 2013; Aalen et al., 2016), and stochastic grammars to learn more inter-pretable structures (Brendel and Todorovic, 2011; Zhou and De la Torre, 2012; Park et al., 2016).”
3D Human Activity Recognition with Reconfigurable Convolutional Neural Networks (2014, background): “re studied in [30] to represent 3D human activities with a new feature called local occupancy pattern (LOP). To handle more complicated activities with large temporal variations, some powerful models [26, 31, 2] further discovered temporal structures of activities by localizing sequential actions. F”
Structural-RNN: Deep Learning on Spatio-Temporal Graphs (2015, methodology): “Spatio-temporal graphs (st-graphs) are a popular [39, 37, 4, 11, 32, 65, 22] general tool for representing such high-level spatio-temporal structures.”
Factorized Graph Matching (2016, methodology): “Our algorithm exploits the particular structure of the afﬁnity matrix and is able to factorize it as a Kronecker product of four smaller matrices.”
Spatiotemporal Deformable Part Models for Action Detection (2013, methodology): “We believe that a hybrid action detection system that incorporates our ideas could achieve further gains.”
Video Action Detection with Relational Dynamic-Poselets (2014, background): “Several previous works [9,3,19,28,18] have considered the relations among parts for action recognition and detection.”
A Deep Structured Model with Radius–Margin Bound for 3D Human Activity Recognition (2015, background): “To handle more complicated activities with large temporal variations, some improved models [38, 43, 3] discovered temporal structures of activities by localizing sequential actions.”
Discriminative prototype selection methods for graph embedding (2013, background): “…Switzerland a r t i c l e i n f o Article history: Received 6 July 2012 Received in revised form 24 September 2012 Accepted 21 November 2012 Keywords: Graph embedding Discriminative prototype selection Graph classification Dissimilarity representation 03/$ - see front matter & 2012 Elsevier Ltd.”
Poselet Key-Framing: A Model for Human Activity Recognition (2013, background): “Therefore, holistic BoW representations of the video perform poorly [26] relative to high-level description-based methods [4, 14, 34] (Table 1).”
Max-Margin Early Event Detectors (2012, background): “2011; Reddy and Shah 2012), and activity recognition (e.g., Brand et al. 1997; Yacoob and Black 1999; Chomat and Crowley 1999; Dollár et al. 2005; Tran and Davis 2008; Nguyen et al. 2009; Ryoo and Aggarwal 2009; Brendel and Todorovic 2011).”
Trajectory-Based Modeling of Human Actions with Motion Reference Points (2012, methodology): “On Olympic Sports, we attain better results than all the state of the arts, including an attribute-based action learningmethod [35] and a graph-based actionmodeling approach [36].”
Discovering discriminative action parts from mid-level video representations (2012, methodology): “Recently, Brendel and Todorovic [5] and Lan et al. [16] proposed activity models that also enable them to learn the relevant action parts of the video.”
Human Action Recognition in Unconstrained Videos by Explicit Motion Modeling (2015, result): “On Olympic Sports, we attain better results than most of the compared approaches, including an attribute-based action learning method [19], a graph-based action modeling approach [61], a new sparse coding-based representation [62], a method modeling the dynamics of action attributes over time [64], ”
Fast Exact Hyper-graph Matching with Dynamic Programming for Spatio-temporal Data (2014, unknown): “In a different vein, Brendel and Todorovic [10] formulated the graph matching problem as a weighted least squares problem on the set permutation matrices.”
Representing Videos Using Mid-level Discriminative Patches (2013, background): “These examples show how strong correspondence can allow us to perform tasks such as object detection, pose estimation and predicting temporal extent.”
Prediction of Human Activity by Discovering Temporal Sequence Patterns (2014, background): “an effective representation for prediction, human activities can be characterized by a complex temporal composition of constituent simple actions and interacting objects.”
Activity representation with motion hierarchies (2013, unknown): “In addition, we use neither video segmentation techniques [Brendel and Todorovic, 2011, Brox and Malik, 2010, Grundmann et al, 2008, Lezama et al, 2011], nor manually pre-deﬁned semantic attribute detectors [Prest et al, 2012, Sadanand and Corso, 2012].”
Matching Guided Distillation (2020, unknown): “recognition [1].”