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 neural network-based novelty detector for image sequence analysis.

Markos Markou1, Sameer Singh

  • 1GORDIOU DESMOU 35, 6045 Larnaca, Cyprus. m.markou@exeter.ac.uk

IEEE Transactions on Pattern Analysis and Machine Intelligence
|September 22, 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

Sex Differences in Arterial Grafting in Patients Undergoing Dialysis and Coronary Artery Bypass Grafting.

JAMA surgery·2026
Same author

Internal thoracic artery grafting confers a survival benefit even in the most elderly population undergoing coronary artery bypass grafting.

JTCVS open·2025
Same author

Arterial Grafting Survival Benefit in Patients With End-Stage Renal Disease Undergoing Coronary Artery Bypass Grafting.

The Annals of thoracic surgery·2025
Same author

Airport microgrid control using integral-of-absolute-error having high renewable penetration.

Scientific reports·2025
Same author

Geographic Variation Exists in Heart Transplantation for Status One and Two Patients After the 2018 Heart Allocation Policy Change.

Clinical transplantation·2025
Same author

Chemical Protein Engineering: Backbone Cyclization Rescues Folding of a 183-Residue Truncated Domain of Malaria Parasite Protein PfAMA1.

Chemistry (Weinheim an der Bergstrasse, Germany)·2025
Same journal

TraGraph-GS: Trajectory Graph-based Gaussian Splatting for Arbitrary Large-Scale Scene Rendering.

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

SWIFT: A Small-World Interaction Framework for Flow-Aware Trajectory Prediction in Autonomous Driving.

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

This study introduces a novel neural network model for image sequence novelty detection. The optimized detector achieves performance comparable to an ideal detector, demonstrating its effectiveness in identifying new patterns.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Artificial Intelligence

Background:

  • Traditional novelty detection methods struggle with complex image sequences.
  • The need for robust systems to identify previously unseen data patterns is critical in various applications.

Purpose of the Study:

  • To propose and evaluate a novel neural network-based model for novelty detection in image sequences.
  • To demonstrate the model's ability to identify and classify novel classes within video data.
  • To compare the proposed model's performance against established baseline novelty detection techniques.

Main Methods:

  • Utilized artificially generated negative data to establish closed decision boundaries with a multilayer perceptron.
  • Implemented a novelty filtering mechanism by thresholding outputs from multiple class-specific neural networks.

Related Experiment Videos

  • Employed clustering to identify novel classes, followed by retraining networks on newly labeled data.

    • Main Results:

    • The proposed novelty detection model successfully identified novel classes in video-based image sequence data.
    • Performance evaluation using two metrics showed the model's superiority over five baseline methods.
    • The optimized novelty detector achieved performance statistically equivalent to an ideal detector (p < 0.05, Chi-square metric).

    Conclusions:

    • The proposed neural network model offers a highly effective approach for novelty detection in image sequences.
    • The method demonstrates significant potential for applications requiring real-time identification of new patterns.
    • Retraining strategies further enhance the adaptability and performance of the novelty detection system.