Locally time-invariant models of human activities using trajectories on the grassmannian

Turaga, Pavan; Chellappa, Rama

doi:10.1109/cvpr.2009.5206710

Cited by 35 publications

(13 citation statements)

References 26 publications

(34 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A human action can generally be considered as a trajectory on a Riemannian manifold M with a Riemannian metric ·, · . For example, at any time instance, the human body can be represented as a shape silhouette, thus a point on the Grassmann manifold [14], or the combination of skeletal joints as a point in a Lie group SE(3) × . .…”

Section: Mathematical Frameworkmentioning

confidence: 99%

Temporal Reflection Symmetry of Human Actions: A Riemannian Analysis

Wang

Turaga

2015

Procedings of the Proceedings of the 1st International Workshop on DIFFerential Geometry in Computer Vision for Analysis of Sha

Self Cite

View full text Add to dashboard Cite

While spatial symmetry of objects is studied frequently, the problem of measuring temporal symmetry has rarely been explored. In this paper we present a mathematical framework that utilizes a recently developed differential geometric approach to quantify temporal reflection symmetry in movements from video data, in which human movements are modeled as trajectories on Riemannian manifolds. Using the recently proposed transport square-root velocity functions (TSRVF) representation, the amounts of temporal asymmetry in movements are decomposed into a spatial component and a temporal one, each of which is quantified by an asymmetry score.

show abstract

Section: Mathematical Frameworkmentioning

confidence: 99%

Temporal Reflection Symmetry of Human Actions: A Riemannian Analysis

Wang

Turaga

2015

Procedings of the Proceedings of the 1st International Workshop on DIFFerential Geometry in Computer Vision for Analysis of Sha

Self Cite

View full text Add to dashboard Cite

show abstract

“…[12] proposes to use distance between linear dynamical systems for action classification. [13], [14] perform action recognition by defining actions as trajectories on the Grasmann Steifel manifold. [15] extends the DTW framework using average templates with multiple features to model intra-class variances and perform simultaneous recognition and localization of actions in a video sequence.…”

Section: Related Workmentioning

confidence: 99%

An Unsupervised Framework for Action Recognition Using Actemes

Kulkarni

Boyer

Horaud

et al. 2011

Computer Vision – ACCV 2010

View full text Add to dashboard Cite

Abstract. In speech recognition, phonemes have demonstrated their efficacy to model the words of a language. While they are well defined for languages, their extension to human actions is not straightforward. In this paper, we study such an extension and propose an unsupervised framework to find phoneme-like units for actions, which we call actemes, using 3D data and without any prior assumptions. To this purpose, build on an earlier proposed framework in speech literature to automatically find actemes in the training data. We experimentally show that actions defined in terms of actemes and actions defined by whole units give similar recognition results. We define actions out of the training set in terms of these actemes to see whether the actemes generalize to unseen actions. The results show that although the acteme definitions of the actions are not always semantically meaningful, they yield optimal recognition accuracy and constitute a promising direction of research for action modeling.

show abstract

“…Furthermore, our model is free from complex and computationally heavy learning and inference on a graph. Trajectory-based representation of activity got encouraging results [14,15,5]. We go further in this direction.…”

Section: Introductionmentioning

confidence: 98%

“…Recent literature has shifted from actions classification in controlled acquisition conditions [1][2][3], to complex activities classification in more realistic scenarios [4][5][6]. In the latter situation, the challenge stems from structured property of activity themselves; specifically, the complicated spatio-temporal relationships between a set of body parts or multiple persons often exhibit low intra-similarity and large inter-variability.…”

Section: Introductionmentioning

confidence: 99%

Middle-Level Representation for Human Activities Recognition: The Role of Spatio-Temporal Relationships

Yuan

Prinet

Yuan

2012

Trends and Topics in Computer Vision

View full text Add to dashboard Cite

Abstract. We tackle the challenging problem of human activity recognition in realistic video sequences. Unlike local features-based methods or global template-based methods, we propose to represent a video sequence by a set of middle-level parts. A part, or component, has consistent spatial structure and consistent motion. We first segment the visual motion patterns and generate a set of middle-level components by clustering keypoints-based trajectories extracted from the video. To further exploit the interdependencies of the moving parts, we then define spatio-temporal relationships between pairwise components. The resulting descriptive middle-level components and pairwise-components thereby catch the essential motion characteristics of human activities. They also give a very compact representation of the video. We apply our framework on popular and challenging video datasets: Weizmann dataset and UT-Interaction dataset. We demonstrate experimentally that our middle-level representation combined with a χ 2 -SVM classifier equals to or outperforms the state-of-the-art results on these dataset.

show abstract

Locally time-invariant models of human activities using trajectories on the grassmannian

Abstract: Human activity analysis is an important problem in com-

Cited by 35 publications

References 26 publications

Temporal Reflection Symmetry of Human Actions: A Riemannian Analysis

Temporal Reflection Symmetry of Human Actions: A Riemannian Analysis

An Unsupervised Framework for Action Recognition Using Actemes

Middle-Level Representation for Human Activities Recognition: The Role of Spatio-Temporal Relationships

Contact Info

Product

Resources

About