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

Experiments of Image Classification Using Dissimilarity Spaces Built with Siamese Networks.

Loris Nanni1, Giovanni Minchio1, Sheryl Brahnam2

  • 1Department of Information Engineering (DEI), Via Gradenigo 6, 35131 Padova, Italy.

Sensors (Basel, Switzerland)
|March 6, 2021
PubMed
Summary

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

Early public childcare and fertility: A longitudinal study for Europe.

PloS one·2026
Same author

Ensemble Deep Learning Models on Raw DNA Sequences for Viral Genome Identification in Human Samples.

Sensors (Basel, Switzerland)·2026
Same author

Deprescribing antihypertensive medications in older people: a systematic review and a meta-analysis.

BMC geriatrics·2026
Same author

Matrix-based vector representations in neural networks for classifying molecular biology data.

Bioinformatics advances·2025
Same author

Entropy-Regularized Attention for Explainable Histological Classification with Convolutional and Hybrid Models.

Entropy (Basel, Switzerland)·2025
Same author

Advancing Precision: A Comprehensive Review of MRI Segmentation Datasets from BraTS Challenges (2012-2025).

Sensors (Basel, Switzerland)·2025

This study introduces a novel image classification system using Siamese Neural Networks (SNNs) to predict patterns in vector spaces. The method achieves competitive and state-of-the-art results on diverse datasets without ad-hoc optimization.

Area of Science:

  • Computer Science
  • Machine Learning
  • Artificial Intelligence

Background:

  • Traditional image classifiers operate within feature spaces.
  • A new approach is needed for robust pattern prediction in diverse domains.

Purpose of the Study:

  • To develop and evaluate an image classification system that predicts patterns in a vector space.
  • To demonstrate the system's versatility across medical and audio-visual data.

Main Methods:

  • Combines dissimilarity spaces from multiple Siamese Neural Networks (SNNs).
  • Utilizes supervised k-means clustering to identify data centroids.
  • Extracts vector space descriptors by projecting patterns onto similarity spaces.
  • Employs Support Vector Machines (SVMs) for classification based on dissimilarity vectors.
Keywords:
audio sound classificationclusteringdissimilarity spaceimage classificationprototype selectionsiamese network

Related Experiment Videos

Main Results:

  • Achieves competitive performance against existing state-of-the-art methods.
  • Obtains state-of-the-art results on one medical dataset.
  • Demonstrates effectiveness on both medical imaging and animal vocalization spectrograms.
  • Performs well without dataset-specific optimization of clustering.

Conclusions:

  • The proposed SNN-based vector space approach offers a versatile and effective method for image classification.
  • This technique shows promise for applications in medical diagnostics and bioacoustics.
  • The system provides a robust alternative to traditional feature-space classifiers.