Learning latent temporal structure for complex event detection

By Kevin D. Tang, Li Fei-Fei, Daphne Koller. 2012

429 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

3D Human Activity Recognition with Reconfigurable Convolutional Neural Networks (2014, methodology): “[26] captured the latent temporal structures of 2D activities based on the variable-duration hidden Markov model.”
Learning human activities and object affordances from RGB-D videos (2012, methodology): “…work on human activity detection from images (Yang et al., 2010; Yao et al., 2011) and from videos (Laptev et al., 2008; Liu et al., 2009; Hoai et al., 2011; Shi et al., 2011; Matikainen et al., 2012; Pirsiavash and Ramanan, 2012; Rohrbach et al., 2012; Sadanand and Corso, 2012; Tang et al., 2012).”
Actions ~ Transformations (2015, methodology): “Temporal structure of videos have also been shown effective in action recognition [47, 6, 44, 13, 34].”
Every Moment Counts: Dense Detailed Labeling of Actions in Complex Videos (2015, background): “Tang et al. [39] extend HMMs to model the duration of each hidden state in addition to the transition parameters of hidden states.”
Connectionist Temporal Modeling for Weakly Supervised Action Labeling (2016, background): “Our goal of understanding the temporal structure of video is related to [10,31,34,37,39,40].”
Modeling Spatial-Temporal Clues in a Hybrid Deep Learning Framework for Video Classification (2015, neutral): “For instance, Li et al. proposed to replace the hidden states in HMMs with visualizable salient poses estimated by Gaussian Mixture Models [24], and Tang et al. introduced latent variables over video frames to discover the most discriminative states of an event based on a variable duration HMM [39].”
Anticipating Human Activities Using Object Affordances for Reactive Robotic Response (2013, background): “There has been a significant amount of work in detecting human activities from 2D RGB videos [1], [2], [3], from inertial/location sensors [4], and more recently from RGB-D videos [5], [6], [7].”
CliqueCNN: Deep Unsupervised Exemplar Learning (2016, neutral): “e retrieval [7, 26], and object recognition [22, 27, 10, 9]. Learning similarities is also of particular importance for video parsing [24], posture analysis [15], and understanding of complex actions [31, 2, 1]. Among the many approaches for similarity learning, supervised techniques have been parti”
NeuralNetwork-Viterbi: A Framework for Weakly Supervised Video Learning (2018, background): “Most existing work on temporal action segmentation focuses on a fully supervised task [15, 26, 31, 19, 32, 22, 6, 37, 18, 36, 28].”
Asynchronous Temporal Fields for Action Recognition (2016, background): “Capturing the whole information of the video in terms of temporal evolution of the video stream has been the focus of some recent approaches [51, 6, 12, 35, 49, 30].”
Bilinear heterogeneous information machine for RGB-D action recognition (2015, methodology): “Related Work Previous action recognition approaches mainly focus on RGB action videos [9, 15, 17, 6].”
ActionVLAD: Learning Spatio-Temporal Aggregation for Action Classification (2017, background): “Other works in event detection and action classification rely on pooling hand crafted features over segments of videos [17, 30, 32, 37, 48].”
Learning latent spatio-temporal compositional model for human action recognition (2013, methodology): “A lot of works were proposed to build temporal structure models [9, 25, 35, 4] based on discriminative and interesting motion segments of the video.”
Compositional Models for Video Event Detection: A Multiple Kernel Learning Latent Variable Approach (2013, methodology): “We use mean average precision (mAP) as the performance metric to remain comparable with recently published works [1, 12].”
Learning latent structure for activity recognition (2014, methodology): “Other work [14] augments an additional layer of latent variables to the linear-chain CRFs.”
Weakly Supervised Action Labeling in Videos under Ordering Constraints (2014, background): “latent action parts [26] or hidden Markov models [33].”
MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation (2019, methodology): “A hidden Markov model is also used in [26] to model both transitions between states and their durations.”
The THUMOS challenge on action recognition for videos "in the wild" (2016, neutral): “…developed for classiﬁcation ( Bojanowski et al., 2014; Duchenne et al., 2009; Karpathy et al., 2014; Niebles et al., 2010; Raptis and Sigal, 2013; Tang et al., 2012 ) and localization ( Hoai et al., 2011; Ke et al., 20 05; 20 07; Pirsiavash and Ramanan, 2014; Tian et al., 2013; Yuan et al.,…”
A Comparative Study of Encoding, Pooling and Normalization Methods for Action Recognition (2012, methodology): “The classical approaches to obtain BoVW representations of action video usually consists of three steps: feature extraction, vector quantilization (VQ) of feature, and sum-pooling with 1 normalization to construct histogram representation [5,9,10,11,12].”