Representation and optimal recognition of human activities

Published 7 November 2002

Somboon Hongeng, François Brémond, R. Nevatia

DOI10.1109/cvpr.2000.855905

Citations80

TL;DR

Towards the goal of realizing a generic automatic human activity recognition system, a new formalism is proposed using a chained hierarchical representation using three type of entities: image features, mobile object properties and scenarios.

Abstract

Towards the goal of realizing a generic automatic human activity recognition system, a new formalism is proposed. Activities are described by a chained hierarchical representation using three type of entities: image features, mobile object properties and scenarios. Taking image features of tracked moving regions from an image sequence as input, mobile object properties are first computed by specific methods while noise is suppressed by statistical methods. Scenarios are recognized from mobile object properties based on Bayesian analysis. Several scenarios are recognized by an algorithm using a probabilistic finite-state automaton (a variant of structured HMM). A demonstration of the optimality of this recognition method is discussed. Finally, the validity and the effectiveness of our approach is demonstrated on both real-world and perturbed data.

Keywords

Computer Science

Coupled hidden Markov models for complex action recognition

1,059 Citations2002Matthew Brand, Nuria Oliver +1 more

Algorithms for coupling and training hidden Markov models (HMMs) to model interacting processes, and demonstrate their superiority to conventional HMMs in a vision task classifying two-handed actions are presented.

Real-time American Sign Language recognition from video using hidden Markov models

770 Citations2002Thad Starner, Alex Pentland

A real-time HMM-based system for recognizing sentence level American Sign Language (ASL) which attains a word accuracy of 99.2% without explicitly modeling the fingers.

Machine LearningAdaptive Probabilistic Networks with Hidden Variables

381 Citations1997John Binder, Daphne Koller +2 more

This paper presents a gradient-based algorithm and shows that the gradient can be computed locally, using information that is available as a byproduct of standard inference algorithms for probabilistic networks.

View PDF

A framework for recognizing multi-agent action from visual evidence

168 Citations1999Stephen Intille, Aaron Bobick

The current work on recognizing American football plays from noisy trajectory data is described, and a probabilistic framework for representing and visually recognizing complex multi-agent action is presented.

Artificial IntelligenceVisual surveillance in a dynamic and uncertain world

165 Citations1995Hilary Buxton, Shaogang Gong

It is demonstrated that using expectations of object trajectory, size and speed for the particular scene improves robustness and sensitivity in dynamic tracking and segmentation and behavioral evaluation under attentional control using a combination of a static BBN TASKNET and dynamic network.

Action recognition using probabilistic parsing

148 Citations2002Aaron Bobick, Yuri Ivanov

A new approach to the recognition of temporal behaviours and activities is presented that provides techniques for generating a discrete symbol stream from continuous low level detectors and for enforcing temporal exclusion constraints during parsing.

Recognition and interpretation of parametric gesture

95 Citations2002Andrew D. Wilson, Aaron Bobick

The approach is to extend the standard hidden Markov model method of gesture recognition by including a global parametric variation in the output probabilities of the states of the HMM.

A Bayesian approach to human activity recognition

91 Citations2003Anant Madabhushi, J.K. Aggarwal

A new approach to human activity recognition that incorporates a Bayesian framework is presented, which finds application in environments where constant monitoring of human activity is required, such as in department stores and airports.

Scenario Recognition in Airborne Video Imagery

34 Citations2003Francois Br emond, Gérard Medioni

A system for image sequence understanding that represents a scenario describing human activities as a combination of sub-scenarios and properties on the mobile objects involved in the scenario and uses automata as methods to recognize scenarios.