Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jun 26, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

Unsupervised activity perception in crowded and complicated scenes using hierarchical bayesian models.

Xiaogang Wang¹, Xiaoxu Ma, W E L Grimson

¹MIT, Cambridge, MA 02139, USA. xgwang@csail.mit.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence

|January 17, 2009

Summary

This study introduces an unsupervised learning framework using hierarchical Bayesian models to identify activities and interactions in complex scenes. The method models visual features, atomic activities, and interactions without human labeling, enabling advanced visual surveillance tasks.

Related Concept Videos

Parallel Processing

Parallel Processing

The brain processes sensory information rapidly due to parallel processing, which involves sending data across multiple neural pathways at the same time. This method allows the brain to manage various sensory qualities, such as shapes, colors, movements, and locations, all concurrently. For instance, when observing a forest landscape, the brain simultaneously processes the movement of leaves, the shapes of trees, the depth between them, and the various shades of green. This enables a quick and...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

FOXF1 promotes ovarian cancer metastasis by facilitating HMGA2-mediated USP30-dependent S100A6 deubiquitination.

Biochimica et biophysica acta. Molecular basis of disease·2024

Same author

Quinazolinone-to-Isoquinoline Metamorphosis by Ruthenium-Catalyzed [4+2] Annulation with Sulfoxonium Ylides.

Organic letters·2024

Same author

Synthesis of Functionalized Indoles by an Iridium-Catalyzed N-H Insertion Cascade: Nucleophilic Cyclization of Naphthylamines with α-Diazocarbonyl Compounds.

The Journal of organic chemistry·2024

Same author

CLIP170 enhancing FOSL1 expression via attenuating ubiquitin-mediated degradation of β-catenin drives renal cell carcinoma progression.

Cellular and molecular life sciences : CMLS·2024

Same author

Derangetropy in Probability Distributions and Information Dynamics.

Entropy (Basel, Switzerland)·2024

Same author

AI-based 3D analysis of retinal vasculature associated with retinal diseases using OCT angiography.

Biomedical optics express·2024

Area of Science:

Computer Vision
Machine Learning
Artificial Intelligence

Background:

Modeling complex activities and interactions in crowded visual scenes is challenging.
Existing methods often require significant human labeling or tracking, limiting scalability.

Purpose of the Study:

To develop a novel unsupervised learning framework for activity and interaction modeling in complex visual surveillance.
To advance existing language models for visual data analysis.

Main Methods:

Utilized hierarchical Bayesian models to link low-level visual features, atomic activities, and multi-agent interactions.
Proposed three specific models: Latent Dirichlet Allocation (LDA) mixture model, Hierarchical Dirichlet Process (HDP) mixture model, and Dual Hierarchical Dirichlet Processes (Dual-HDP).

Related Experiment Videos

Last Updated: Jun 26, 2026

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

Employed unsupervised learning to cluster moving pixels into activities and video clips into interactions.

Main Results:

Successfully discovered typical atomic activities and interactions in challenging datasets (traffic, train stations).
Demonstrated effective segmentation of long video sequences into interactions and motions into activities.
Achieved abnormality detection and supported high-level queries on activities and interactions without human supervision.

Conclusions:

The proposed unsupervised framework offers a robust solution for visual surveillance in crowded environments.
Hierarchical Bayesian models provide a powerful approach to understanding complex activities and interactions from visual data.
The framework eliminates the need for manual tracking and labeling, enhancing efficiency and applicability.