Local Trinary Patterns for human action recognition

By Lahav Yeffet, Lior Wolf. IEEE International Conference on Computer Vision, 2009

379 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/68080fa24fef0f6eaa3dfd6f63978de01bc251bf

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

A database for fine grained activity detection of cooking activities (2012, background): “To create a discriminative feature representation of a video many approaches first detect spacetime interest points [6, 15] or sample them densely [39] and then extract diverse descriptors in the image-time volume, such as histograms of oriented gradients (HOG) and flow (HOF) [16] or local trinary p”
Dense Trajectories and Motion Boundary Descriptors for Action Recognition (2013, neutral): “Feature descriptors range from higher order derivatives (lo-cal jets), gradient information, optical ﬂow, and brightness information (Doll´ar et al, 2005; Laptev et al, 2008; Sch¨uldt et al, 2004) to spatio-temporal extensions of image descriptors, such as 3D-SIFT (Scovanner et al, 2007), HOG3D (Kl¨”
Discriminative figure-centric models for joint action localization and recognition (2011, methodology): “Reported results [19, 22, 24, 13, 12] on this dataset use Leave-One-Out (LOO) cross validation, cycling each example as a test video one at a time.”
Evaluating New Variants of Motion Interchange Patterns (2013, background): “The Local Trinary Patterns (LTP) descriptor of [31] is another LBP extension to videos.”
A Hough transform-based voting framework for action recognition (2010, result): “We outperform [20] and [36], and have comparable results with [34] though the we use much simpler features (their best results were achieved using 3D-HOG descriptors).”
Multi-task Sparse Learning with Beta Process Prior for Action Recognition (2013, methodology): “The UCF sports database was tested in a leave-one-out manner, cycling each example in as a test video one at a time, following [18] [24] [22].”
The Action Similarity Labeling Challenge (2012, background): “These sets, although new, have already drawn a lot of attention (see, for example, [13], [26], [27]).”
Real-Time Hand Gesture Detection and Recognition Using Bag-of-Features and Support Vector Machine Techniques (2011, neutral): “They permit to recognize a rich set of actions ranging from simple periodic motion (waving, running) to interactions (kissing, shaking hands) [45]–[48].”
Multi-Stream Deep Neural Networks for RGB-D Egocentric Action Recognition (2019, methodology): “on the first two video-based datasets, due to its better performance compared with the other exsiting hand-crafted spatiotemporal descriptors [33], [98].”
Hough Forests for Object Detection, Tracking, and Action Recognition (2011, result): “[80] and [87] and have comparable results with [32] and”
Action bank: A high-level representation of activity in video (2012, background): “Introduced by [18], sparse space-time interest points and subsequent methods, such as local trinary patterns [39], dense interest points [37], page-rank features [24], and discriminative class-specific features [16], typically compute a bag of words representation on local features and sometimes loc”
A Cause and Effect Analysis of Motion Trajectories for Modeling Actions (2014, background): “Often local features for the interest points are based on gradient information, optical flow [4, 15, 22, 27], local trinary patterns [31], 3D-SIFT [23].”
Real-time Action Recognition by Spatiotemporal Semantic and Structural Forests (2010, methodology): “For instance, Yeffet and Wolf [29] utilise dense local trinary patterns with a linear SVM classifier.”
Discovering discriminative action parts from mid-level video representations (2012, methodology): “We note here that earlier studies [23, 34] have used a leave-one-out cross validation setup, which retains the above correlation.”
A temporal Bayesian model for classifying, detecting and localizing activities in video sequences (2012, neutral): “Various other descriptors such as those based on HoG/HoF) [15], HoG3D [11], 3D-SIFT [20] and Local Trinary Patterns [28] have also been proposed to describe interest points.”
Action recognition by dense trajectories (2011, background): “Experimental results show a significant improvement over the state of the art on four datasets of varying difficulty, i.e. KTH, YouTube, Hollywood2 and UCF sports.”
Learning Sparse Representations for Human Action Recognition (2012, background): “2 Yeffet & Wolf [29] 79.”