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

Image transform bootstrapping and its applications to semantic scene classification.

Jiebo Luo, Matthew Boutell, Robert T Gray

    IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
    |June 24, 2005
    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

    Surgical outcomes in gallbladder cancer: evidence from the UK nationwide CAPBIL study.

    HPB : the official journal of the International Hepato Pancreato Biliary Association·2026
    Same author

    The first-in-human ENCIT01 trial comparing second- versus third-generation L1CAM-specific CAR T cells in patients with primary refractory or relapsed neuroblastoma.

    Clinical cancer research : an official journal of the American Association for Cancer Research·2026
    Same author

    Latent Chain-of-Thought for Visual Reasoning.

    Advances in neural information processing systems·2026
    Same author

    Large Differences in Pulsatility and Resistance of the Middle Cerebral Artery Assessed With Transcranial Doppler Versus Transcranial Color-Coded Doppler: Impact of Post-Processing Approaches.

    Ultrasound in medicine & biology·2026
    Same author

    Management of incidental gallbladder cancer in a nationwide CAPBIL study.

    The British journal of surgery·2026
    Same author

    Acoustic Salience Drives Pupillary Dynamics in an Interrupted, Reverberant Task.

    bioRxiv : the preprint server for biology·2026

    This study introduces image-transform bootstrapping to improve scene classification systems. This novel method enhances training and testing data, boosting performance in image recognition tasks.

    Area of Science:

    • Computer Science
    • Artificial Intelligence
    • Machine Learning

    Background:

    • Exemplar-based scene classification performance relies heavily on training data quality and size.
    • Real-world datasets often contain variations and distracting elements that hinder accurate image matching.

    Purpose of the Study:

    • To introduce a novel image-transform bootstrapping technique to enhance exemplar-based scene classification.
    • To address limitations in training data and improve robustness against variations in testing images.

    Main Methods:

    • Developed image-transform bootstrapping, a method operating in the image space, distinct from traditional feature-space boosting.
    • Proposed three schemes for augmenting training data, testing data, or both using image transforms.
    • Applied the technique to sunset detection, outdoor scene classification, and automatic image orientation detection.

    Related Experiment Videos

    Main Results:

    • Demonstrated successful application across three diverse and increasingly complex image recognition tasks.
    • Showcased the effectiveness of image-transform bootstrapping in improving classification performance.
    • Validated that performance gains are achievable by selecting appropriate transforms and meta-classification methods.

    Conclusions:

    • Image-transform bootstrapping offers a powerful approach to overcome data limitations in exemplar-based systems.
    • The method's adaptability allows for tailored solutions based on specific problem domains and chosen classifiers.
    • This technique significantly enhances the performance and reliability of scene classification systems.