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

Low dimensional adaptive texture feature vectors from class distance and class difference matrices.

Birgitte Nielsen1, Fritz Albregtsen, Håvard E Danielsen

  • 1Department of Informatics, University of Oslo, P.O. Box 1080 Blindern, N-03 16 Oslo, Norway. birgitn@ifi.uio.no

IEEE Transactions on Medical Imaging
|January 15, 2004
PubMed
Summary

This study introduces adaptive texture features for improved classification. New features derived from class distance and difference matrices outperform traditional methods, especially in challenging datasets.

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

Applicability of mitotic figure counting by deep learning: a development and pan-cancer validation study.

FEBS open bio·2026
Same author

Generalisation of automatic tumour segmentation in histopathological whole-slide images across multiple cancer types.

NPJ precision oncology·2026
Same author

Transfer Learning with Interpretability: Liver Segmentation in CT and MR using Limited Dataset.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2025
Same author

Prognostic and therapeutic implication of molecular classification including L1CAM expression in high-risk endometrial cancer.

Gynecologic oncology·2024
Same author

Domoic Acid as a Lead for the Discovery of the First Selective Ligand for Kainate Receptor Subtype 5 (GluK5).

Journal of medicinal chemistry·2024
Same author

Deep learning for automated scoring of immunohistochemically stained tumour tissue sections - Validation across tumour types based on patient outcomes.

Heliyon·2024

Area of Science:

  • Image analysis and pattern recognition
  • Computational pathology
  • Biomedical imaging

Background:

  • Traditional texture analysis relies on high-dimensional, nonadaptive feature vectors extracted from statistical matrices.
  • Feature selection in texture analysis faces challenges with limited data and a large feature pool, risking coincidental selections.
  • Understanding texture is crucial for reliable classification and modeling complex phenomena.

Purpose of the Study:

  • To develop a unified approach for statistical texture feature extraction using adaptive low-dimensional feature vectors.
  • To enhance texture classification reliability and improve the understanding of modeled phenomena.
  • To evaluate the performance of novel adaptive features against classical methods.

Main Methods:

Related Experiment Videos

  • Utilized class distance and class difference matrices for adaptive feature vector generation.
  • Applied the unified approach to four established texture analysis methods.
  • Tested the adaptive features on a challenging dataset of 45 Brodatz texture pairs and ovarian cancer cell nucleus images.

    • Main Results:

    • The novel adaptive features demonstrated superior performance compared to classical features on difficult texture datasets.
    • Class distance and difference matrices effectively distinguished between texture patterns in ovarian cancer cell nuclei from different prognostic classes.
    • A single adaptive feature often captured the majority of the discriminatory power for each tested texture analysis method.

    Conclusions:

    • Adaptive texture features derived from class distance and difference matrices offer a more robust and efficient approach to texture classification.
    • This method enhances the understanding of texture differences, particularly in biomedical applications like cancer prognostics.
    • The findings suggest that a low-dimensional set of adaptive features can be highly effective for texture analysis.