Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A system for learning statistical motion patterns.

Weiming Hu1, Xuejuan Xiao, Zhouyu Fu

  • 1National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, PO Box 2728, Beijing 100080, P.R. China. wmhu@nlpr.ia.ac.cn

IEEE Transactions on Pattern Analysis and Machine Intelligence
|August 26, 2006
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Chemotherapy With or Without Anti-EGFR Agents in Left- and Right-Sided Metastatic Colorectal Cancer: An Updated Meta-Analysis.

Journal of the National Comprehensive Cancer Network : JNCCN·2019
Same author

Zscan4c activates endogenous retrovirus MERVL and cleavage embryo genes.

Nucleic acids research·2019
Same author

Kinesin family member C1 accelerates bladder cancer cell proliferation and induces epithelial-mesenchymal transition via Akt/GSK3β signaling.

Cancer science·2019
Same author

MSI2-TGF-β/TGF-β R1/SMAD3 positive feedback regulation in glioblastoma.

Cancer chemotherapy and pharmacology·2019
Same author

METTL3 facilitates tumor progression via an m<sup>6</sup>A-IGF2BP2-dependent mechanism in colorectal carcinoma.

Molecular cancer·2019
Same author

Association between suicide and multiple sclerosis: An updated meta-analysis.

Multiple sclerosis and related disorders·2019
Same journal

HardFlow: Hard-Constrained Sampling for Flow-Matching Models Via Trajectory Optimization.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Industrial Brain: Self-Evolving Neuro-Symbolic Autonomy with Causal Resilience for Cyber-Physical Systems.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Adaptive Hardness-Driven Dictionary Distillation for Incomplete Streaming View Clustering.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Task-KV: Task-aware KV Cache Optimization via Semantic Differentiation of Attention Heads.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Achieving Text-based Person Retrieval with Any Granularity.

IEEE transactions on pattern analysis and machine intelligence·2026
See all related articles

This study introduces an automated system for learning object motion patterns to detect anomalies and predict behavior. The approach robustly tracks objects and statistically models their movements for improved scene understanding.

Area of Science:

  • Computer Vision
  • Artificial Intelligence
  • Machine Learning

Background:

  • Traditional motion pattern analysis relies on predefined scene knowledge.
  • Automatic construction of scene-specific motion patterns is highly desirable.
  • Existing methods lack robustness in complex, dynamic environments.

Purpose of the Study:

  • To develop a system for automatically learning motion patterns.
  • To enable robust multi-object tracking for anomaly detection and behavior prediction.
  • To create a data-driven approach for understanding dynamic scenes.

Main Methods:

  • A robust multi-object tracking algorithm using fuzzy K-means clustering for foreground pixels.
  • Hierarchical clustering of object trajectories based on spatial and temporal information.

Related Experiment Videos

  • Representation of motion patterns using chains of Gaussian distributions.
  • Statistical methods for anomaly detection and behavior prediction based on learned patterns.

    • Main Results:

    • Demonstrated robustness of the multi-object tracking algorithm.
    • Showcased the efficiency of the motion pattern learning algorithm.
    • Achieved encouraging performance in anomaly detection and behavior prediction.
    • Validated the system on real and model traffic scene image sequences.

    Conclusions:

    • The proposed system effectively learns motion patterns for anomaly detection and behavior prediction.
    • The automated approach enhances understanding of dynamic scenes without prior knowledge.
    • The system offers a robust and efficient solution for analyzing complex motion data.