Chaotic Invariants for Human Action Recognition

By Saad Ali, Arslan Basharat, Mubarak Shah. IEEE International Conference on Computer Vision, 2007

329 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistical and Nonlinear Physics, Signal Processing

View Paper: https://www.semanticscholar.org/paper/168b92659ead59f0a7733a6594f3249817f469b9

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 Statistical and Nonlinear Physics

In Signal Processing

Citations in the corpus

Expandable Data-Driven Graphical Modeling of Human Actions Based on Salient Postures (2008, methodology): “As adopted in most previous works [3], [12], [26], [27] using the same data set, we conducted leave-one-sample-out test to verify the overall performance of the proposed model.”
Human Action Recognition With Video Data: Research and Evaluation Challenges (2014, methodology): “[104] × × S × × × × √ [81], [85], [91]–[97], [108], [113], [127], [153], [154] × × I I × × √ ×”
A Cause and Effect Analysis of Motion Trajectories for Modeling Actions (2014, methodology): “Similar to the proposed work, few of the trajectory-based methods to perform action classification are presented in [1, 17, 29, 25, 18, 11].”
Pose primitive based human action recognition in videos or still images (2008, unknown): “[2] 89.”
Action snippets: How many frames does human action recognition require? (2008, methodology): “Past research in this domain can be roughly classified into two approaches: one that extracts a global feature set from a video [1, 9, 18, 27], and aims to assign a single label to the entire video, using these features.”
Chaotic invariants of Lagrangian particle trajectories for anomaly detection in crowded scenes (2010, methodology): “Based on the Taken embedding principle (see Section 3 in [7] for preliminaries of chaotic embedding), we are able to reconstruct a strange attractor in phase space that has equivalent dynamics to the system revealed by time series.”
Cross-View Action Recognition from Temporal Self-similarities (2008, methodology): “6% achieved in by a recent trajectory-based method in [17].”
Exploring discriminative pose sub-patterns for effective action classification (2013, background): “Similar methodologies can be found in [26, 35, 28, 3], where either trajectories formed by joints [28, 35, 3] or curvature patterns of trajectories formed from body parts [26] are treated as features.”
Moving vistas: Exploiting motion for describing scenes (2010, background): “Here, we show how the motion-attributes discussed in section 2 can be used to provide semantically meaningful organization of videos.”
Action bank: A high-level representation of activity in video (2012, background): “, [2, 29], which itself is challenging and unsolved.”
Action recognition by learning discriminative key poses (2011, background): “Also, unlike various other approaches [21, 1], we do not require any subsampling, upsampling or trimming during training.”
Recognizing human actions using multiple features (2008, background): “Recently, [10] used a series of 3D poses to represent the shape of the actor and [4] discovered the chaotic invariant features from the joint tracking for action recognition.”
Histograms of oriented optical flow and Binet-Cauchy kernels on nonlinear dynamical systems for the recognition of human actions (2009, methodology): “in [2] used joint trajectories to extract invariant features to model the non-linear dynamics of human activities.”
Local Trinary Patterns for human action recognition (2009, result): “, [1]), and is in contrast with space-time gradient methods (e.”
Trajectons: Action recognition through the motion analysis of tracked features (2009, methodology): “…trajectons, each trajectory snippet is simply a concatenated vector of (dx, dy) derivatives for the trajectory, whereas in AffineAugmented (AA) trajectons, this vector of derivatives is concatenated with a vector of local affine transforms to describe the motion around each trajectory point.”
Discriminative subvolume search for efficient action detection (2009, background): “One type of approaches uses motion trajectories to represent actions and it requires target tracking [16] [1].”
Sequence of the Most Informative Joints (SMIJ): A new representation for human skeletal action recognition (2012, methodology): “…3D positions of joints or joint angle configurations , most of the methods use linear dynamical systems (LDS) or non-linear dynamical systems (NLDS), e.g., in [8] [9] [10], or hidden Markov models (HMM), see, e.g., the earlier work by Yamato et al. [11] and a review of several others in [12].”
Slow Feature Analysis for Human Action Recognition (2012, methodology): “Laptev and Lindeberg [11] extended the 2D spatial interest points to detect 3D spatiotemporal interest points for action representation.”