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

Selecting salient frames for spatiotemporal video modeling and segmentation.

Xiaomu Song1, Guoliang Fan

  • 1Evanston Northwestern Healthcare Research Institute and Northwestern University, Evanston, IL 60201, USA. xiaomu-song@northwestern.edu

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|December 21, 2007
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

Render-Rank-Refine: Accurate 6D Indoor Localization via Circular Rendering.

Journal of imaging·2026
Same author

The Anti-Lupus Plate: Mapping Nutritional Interventions to Inflammatory Pathways in Systemic Lupus Erythematosus.

Food science & nutrition·2025
Same author

[Impact of mean perfusion pressure on the risk of sepsis-associated acute kidney injury].

Zhonghua wei zhong bing ji jiu yi xue·2025
Same author

The application of the triglyceride-glucose-body mass index (TyG-BMI) in predicting acute kidney injury in diabetic patients following coronary artery bypass grafting surgery.

Journal of cardiothoracic surgery·2025
Same author

Process and systematic study of gold recovery from flexible printed circuit boards (FPCBs) using a choline chloride-ethylene glycol system.

Waste management (New York, N.Y.)·2024
Same author

Indoor Camera Pose Estimation from Room Layouts and Image Outer Corners.

IEEE transactions on multimedia·2023
Same journal

Change-Prior-Guided Unsupervised Change Detection of Heterogeneous Remote Sensing Images.

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

AgonicDreamer: Enhancing Multi-View Consistency in Text-to-3D Generation via Rectified Score Distillation.

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

BiCM-Prompt: Bidirectional Cross-Modal Prompt Tuning for Class-Incremental Learning on Multisource Remote Sensing Images.

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

GoP-based Quality Enhancement on Video Compression.

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

Align then Tensorize: Multi-Level Consistent Anchor Graph Learning for Scalable Multi-View Clustering.

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

Beyond Fidelity: Diverse Image Synthesis via Retrieval-Augmented Diffusion.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
See all related articles

We developed a new statistical model for spatiotemporal video segmentation using Gaussian mixture models (GMMs) and frame saliency. This approach efficiently segments videos and can identify object behaviors for tasks like key-frame extraction.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Video Analysis

Background:

  • Spatiotemporal video segmentation is crucial for understanding video content.
  • Existing methods often struggle with data redundancy and irrelevance in video sequences.
  • Gaussian Mixture Models (GMMs) provide a probabilistic framework for modeling data distributions.

Purpose of the Study:

  • To propose a novel statistical generative model for spatiotemporal video segmentation.
  • To introduce frame saliency for efficient video modeling and segmentation.
  • To enable building blocks for semantic video segmentation.

Main Methods:

  • Utilized a six-dimensional spatiotemporal feature space for video modeling.
  • Introduced frame saliency to quantify frame relevance for GMM-based modeling.

Related Experiment Videos

  • Developed a modified expectation-maximization algorithm for simultaneous GMM training and saliency estimation.
  • Refined GMM estimation using extracted salient frames for segmentation.

    • Main Results:

    • The proposed method effectively segments video sequences into homogeneous segments.
    • Frame saliency estimation aids in reducing data redundancy and irrelevance during model training.
    • The method demonstrates effectiveness and efficiency on real-world video data.
    • Identified that frame saliency can imply object behaviors.

    Conclusions:

    • The new statistical generative model offers an effective and efficient approach to spatiotemporal video segmentation.
    • Frame saliency is a valuable concept for improving GMM-based video modeling and segmentation.
    • The method's applicability extends to other video analysis tasks like key-frame extraction and video skimming.