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 wrapper-based approach to image segmentation and classification.

Michael E Farmer1, Anil K Jain

  • 1Department of Computer Sciences, Engineering Science and Physics, University of Michigan, Flint, MI 48502, USA. farmerme@umflint.edu

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|December 24, 2005
PubMed
Summary

This study introduces a novel wrapper method for image segmentation and classification in computer vision. It integrates classification accuracy as a metric to improve segmentation quality, enhancing object recognition in complex scenes.

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

CoDA: Color Distribution Probing for Efficient and Generalizable AI-Generated Image Detection.

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

CT/USG guided biopsy in spinal tuberculosis: Diagnostic yield.

Journal of clinical orthopaedics and trauma·2026
Same author

Early Outcome of Intracapsular Fracture Neck of Femur Managed with Femoral Neck System Versus Cannulated Cancellous Screws: A Prospective Comparative Study.

Indian journal of orthopaedics·2025
Same author

Universal Fingerprint Generation: Controllable Diffusion Model With Multimodal Conditions.

IEEE transactions on pattern analysis and machine intelligence·2024
Same author

Assessment of change in antimicrobial susceptibility pattern of bacteria causing orthopaedic infections by sequential sampling: A prospective cohort study.

Journal of clinical orthopaedics and trauma·2024
Same author

Drug-Resistant Bone, Joint and Spine Tuberculosis: Evolution of Diagnosis and Treatment.

Indian journal of orthopaedics·2024

Area of Science:

  • Computer Vision
  • Machine Learning
  • Image Processing

Background:

  • Traditional segmentation-classification pipelines struggle with reliable object extraction without prior knowledge.
  • Lack of clear metrics hinders segmentation quality evaluation and algorithm comparison.
  • Existing methods often require objects to be homogeneous in low-level image parameters.

Purpose of the Study:

  • To develop an integrated segmentation and classification method addressing segmentation reliability and evaluation.
  • To propose a new paradigm using classification accuracy as a metric for segmentation quality.
  • To create a flexible image segmentation framework adaptable to existing algorithms.

Main Methods:

  • Proposed a wrapper method integrating segmentation and classification, using classification accuracy as the primary metric.

Related Experiment Videos

  • Utilized object classification's contextual information to guide segmentation.
  • Developed a segmentation algorithm relaxing homogeneity requirements by using shape as a classification feature.

    • Main Results:

    • Demonstrated a robust segmentation framework applicable to complex, real-world images.
    • Successfully applied the method to identify infant occupants in vehicles for airbag control.
    • Showcased the potential for broad applicability across diverse real-world scenarios.

    Conclusions:

    • The wrapper-based approach significantly improves segmentation quality and reliability.
    • This framework offers a versatile solution for various image segmentation and classification tasks.
    • The method overcomes limitations of traditional approaches by integrating classification feedback.