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 Video

Updated: Jul 7, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Decision boundary feature extraction for neural networks.

C Lee1, D A Landgrebe

  • 1Sch. of Electr. and Comput. Eng., Purdue Univ., West Lafayette, IN.

IEEE Transactions on Neural Networks
|January 1, 1997
PubMed
Summary

This study introduces a novel feature extraction technique for feedforward neural networks, leveraging decision boundaries to identify essential classification features without assuming data distributions. Experiments demonstrate promising outcomes for this innovative approach.

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

Alveolar cleft bone grafts: results and imprecisions of the dental radiograph.

Plastic and reconstructive surgery·1995
Same author

Correlation between cumulative labeling indices measured in the terminal bronchioles and in the centriacinar region in the lungs of rats exposed to oxidant air pollutants.

Toxicology letters·1995
Same author

Growth suppression of human colorectal carcinoma in nude mice by monoclonal antibody C27-abrin A chain conjugate.

Diseases of the colon and rectum·1995
Same author

Disseminated disease due to Mycobacterium avium complex in AIDS.

QJM : monthly journal of the Association of Physicians·1995
Same author

A general model of invariant chain association with class II major histocompatibility complex proteins.

Proceedings of the National Academy of Sciences of the United States of America·1995
Same author

Postnatal maturation of the blood-brain barrier for unbound bilirubin in newborn piglets.

Brain research·1995

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Machine Learning

Background:

  • Traditional feature extraction methods often rely on assumptions about data distributions.
  • Neural networks excel at modeling complex, arbitrary decision boundaries.
  • Extracting features directly from decision boundaries offers a new paradigm.

Purpose of the Study:

  • To propose a novel feature extraction method for feedforward neural networks.
  • To utilize the decision boundary as the primary source for feature extraction.
  • To develop a method independent of underlying data distribution assumptions.

Main Methods:

  • Defined the decision boundary specifically within the context of neural networks.
  • Developed a procedure to extract all necessary classification features from the defined decision boundary.

Related Experiment Videos

Last Updated: Jul 7, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

  • Built upon the principles of the decision boundary feature extraction algorithm.

    • Main Results:

    • The proposed method successfully extracts features from the neural network's decision boundary.
    • Experimental results indicate the effectiveness and promise of the new feature extraction technique.
    • The approach demonstrated the capability to handle complex problems with arbitrary decision boundaries.

    Conclusions:

    • Feature extraction from decision boundaries is a viable and effective strategy for neural networks.
    • The proposed method offers an alternative to traditional feature extraction, particularly for complex datasets.
    • This work opens new avenues for understanding and utilizing information encoded within neural network decision boundaries.