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 pyramidal neural network for visual pattern recognition.

Son Lam Phung1, Abdesselam Bouzerdoum

  • 1School of Electrical, Computer and Telecommunications Engineering, University of Wollongong, Wollongong, NSW 2522, Australia. s.phung@ieee.org

IEEE Transactions on Neural Networks
|March 28, 2007
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

Single-molecule dynamics of fibrinogen and bovine serum albumin adsorption on silica nanoparticle coatings observed using high-speed atomic force microscopy.

Biointerphases·2025
Same author

Correcting Susceptibility Artifacts of MRI Sensors in Brain Scanning: A 3D Anatomy-Guided Deep Learning Approach.

Sensors (Basel, Switzerland)·2021
Same author

Unbalanced Hybrid AOA/RSSI Localization for Simplified Wireless Sensor Networks.

Sensors (Basel, Switzerland)·2020
Same author

An unsupervised deep learning technique for susceptibility artifact correction in reversed phase-encoding EPI images.

Magnetic resonance imaging·2020
Same author

Compressive Radar Imaging of Stationary Indoor Targets with Low-rank plus Jointly Sparse and Total Variation Regularizations.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2020
Same author

Hybrid Deep Learning-Gaussian Process Network for Pedestrian Lane Detection in Unstructured Scenes.

IEEE transactions on neural networks and learning systems·2020

We introduce PyraNet, a novel neural network architecture inspired by image pyramids and local receptive fields for visual pattern classification. PyraNet demonstrates effective gender determination from facial images, outperforming other methods.

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Machine Learning

Background:

  • Image pyramids and local receptive fields are key concepts in visual pattern recognition.
  • Existing neural network architectures face challenges in efficient feature extraction and dimensionality reduction for complex visual data.

Purpose of the Study:

  • To propose a novel neural network architecture, PyraNet, for visual pattern classification.
  • To investigate the performance of PyraNet using various training methods and error functions.
  • To evaluate PyraNet's effectiveness in gender determination from facial images.

Main Methods:

  • Developed PyraNet, a hierarchical neural network with Pyramidal and 1-D layers.
  • Trained PyraNet using five methods: gradient descent (GD), momentum, resilient back-propagation (RPROP), conjugate gradient (CG), and Levenberg-Marquardt (LM).

Related Experiment Videos

  • Utilized mean-square-error (mse) and cross-entropy (CE) as error functions for training.

    • Main Results:

    • PyraNet successfully performs feature extraction and dimensionality reduction using nonlinear 2-D neurons.
    • Comparative analysis on the FERET database shows PyraNet's performance in gender classification.
    • PyraNet's accuracy is benchmarked against Convolutional Neural Networks (CNN), k-Nearest Neighbors (k-NN), and Support Vector Machines (SVM).

    Conclusions:

    • PyraNet offers a promising new architecture for visual pattern classification tasks.
    • The proposed architecture effectively addresses challenges in feature extraction and dimensionality reduction.
    • PyraNet shows competitive or superior performance in gender determination compared to established classifiers.