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 Unified Framework for Event Summarization and Rare Event Detection from Multiple Views.

Junseok Kwon, Kyoung Mu Lee

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |September 10, 2015
    PubMed
    Summary

    This study introduces a unified graph-based framework for video event summarization and rare event detection. The novel approach effectively identifies key events and anomalies by editing video event graphs, offering improved unsupervised multi-view analysis.

    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

    Particle Diffusion Matching: Random Walk Correspondence Search for the Alignment of Standard and Ultra-Widefield Fundus Images.

    IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
    Same author

    LucidDreamer: Domain-Free Generation of 3D Gaussian Splatting Scenes.

    IEEE transactions on visualization and computer graphics·2025
    Same author

    Active Diffusion Matching: Score-Based Iterative Alignment of Cross-Modal Retinal Images.

    IEEE transactions on bio-medical engineering·2025
    Same author

    FQ-UWF: Unpaired Generative Image Enhancement for Fundus Quality Ultra-Widefield Retinal Images.

    Bioengineering (Basel, Switzerland)·2024
    Same author

    Learning Controllable ISP for Image Enhancement.

    IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2023
    Same author

    AttSec: protein secondary structure prediction by capturing local patterns from attention map.

    BMC bioinformatics·2023

    Area of Science:

    • Computer Vision
    • Machine Learning
    • Artificial Intelligence

    Background:

    • Conventional methods often address event summarization and rare event detection separately.
    • Existing approaches may lack a unified framework for simultaneous analysis.
    • Multi-view video analysis presents challenges in consolidating information.

    Purpose of the Study:

    • To propose a novel, unified framework for event summarization and rare event detection in videos.
    • To develop a graph-based approach that models video events and their relationships.
    • To enable unsupervised, multi-view event analysis and rare transition detection.

    Main Methods:

    • Representing videos as graphs where nodes are events and edges denote relationships.
    • Employing graph editing techniques (subgraph merging, edge pruning) to identify summaries and rare events.

    Related Experiment Videos

  • Utilizing a Markov Chain Monte Carlo (MCMC) method to minimize a custom energy model incorporating causality, frequency, and significance.

    • Main Results:

    • The proposed method accurately generates event summaries for multiple videos in an unsupervised manner.
    • Experimental results demonstrate the approach's effectiveness in detecting rare event transitions.
    • Extension to multi-view scenarios allows for robust event summarization and rare event detection across different perspectives.

    Conclusions:

    • The unified graph editing framework offers a powerful solution for simultaneous event summarization and rare event detection.
    • The method excels in unsupervised learning and provides advantages in identifying infrequent event sequences.
    • The approach is scalable and effective for analyzing complex multi-view video data.