A Hierarchical Model of Shape and Appearance for Human Action Classification

By Juan Carlos Niebles, Li Fei-Fei. 2007 IEEE Conference on Computer Vision and Pattern Recognition, 2007

508 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/7a5fe226444b2ca367ac9ac54b5b4113023fd404

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

Progressive search space reduction for human pose estimation (2008, methodology): “Following the standard leave-one-out evaluation protocol [3, 8, 16], we train on eight subjects and test on the remaining one.”
A Biologically Inspired System for Action Recognition (2007, background): “It has also been suggested [10] that another part of the visual system, the ventral stream of the visual cortex, involved with the analysis of shape may also be important for the recognition of motion (consistent with recent work in computer vision [17] which has shown the benefit of using shape fea”
Human activity prediction: Early recognition of ongoing activities from streaming videos (2011, background): “a single frame [9]) to perform a reliable recognition.”
Advances in View-Invariant Human Motion Analysis: A Review (2010, result): “compared to previous works, this method offers superior classification performance on a number of large human motion datasets [80].”
Video-Based Abnormal Human Behavior Recognition—A Review (2012, background): “Spatial–temporal features have shown particular promise in motion understanding due to its rich descriptive power [76] and,”
Recognizing actions by shape-motion prototype trees (2009, background): “To the best of our knowledge, both of them outperform published results in [1, 5, 8, 9, 16, 18, 21, 23, 24, 30] on the KTH dataset.”
Expandable Data-Driven Graphical Modeling of Human Actions Based on Salient Postures (2008, background): “Researchers have so far explored various types of visual information to describe human motion, including motion trajectories [2]–[4], sequences of silhouettes or contours of the human body [5], [6], spatio–temporal salient points [7], hierarchical configuration of body parts [8], [9], such as torso,”
Machine Recognition of Human Activities: A Survey (2008, background): “Niebles et al. [53] proposed a hierarchical model where the higher level is a constellation of parts much smaller than the actual number of features.”
Human Action Recognition by Semilatent Topic Models (2009, unknown): “Niebles and Fei-Fei [44] combine shape information with local appearance features by building a hierarchical model that can be characterized as a constellation of bag-of-features.”
Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words (2008, background): “richer models that can incorporate geometric information, such as the spatial and temporal arrangement of local features ( Niebles and Fei-Fei 2007 ), as well as explicit models for the human body.”
Action Recognition by Hierarchical Sequence Summarization (2013, background): “Numerous approaches have been proposed to learn from a hierarchical representation of human action [13, 23, 8, 26, 12].”
Learning human actions via information maximization (2008, background): “Spatiotemporal interest points [8, 6, 2, 11 ] have also been widely successful.”
Learning a hierarchy of discriminative space-time neighborhood features for human action recognition (2010, background): “Hierarchical representations for local-feature activity recognition have only been explored to a limited extent [20, 9].”
A combined pose, object, and feature model for action understanding (2012, methodology): “Several works ([20], [15], and [19]) develop this general methodology further to capture the spatial and temporal distribution of these words.”
Local Trinary Patterns for human action recognition (2009, result): “The overall mean accuracy we obtain for this dataset is 79.2%, compared to 69.2% reported in [14].”
Modeling Temporal Structure of Decomposable Motion Segments for Activity Classification (2010, background): “Both generative [17,18,19,20] and discriminative [21,22] models have shown promise in leveraging the spatial structures among parts for object recognition.”
Action snippets: How many frames does human action recognition require? (2008, methodology): “In terms of features used, [10, 17] extract only optic flow, [3, 27] only silhouette shape, and [18] have used either form or motion, while our work, as well as [19], extract both cues independently.”
Modeling Activity-Dependent Plasticity in BCM Spiking Neural Networks With Application to Human Behavior Recognition (2011, unknown): “This paper is concluded by Section VIII.”