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 Video

Updated: Mar 8, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

852

Mesh Convolutional Restricted Boltzmann Machines for Unsupervised Learning of Features With Structure Preservation on

Zhizhong Han, Zhenbao Liu, Junwei Han

    IEEE Transactions on Neural Networks and Learning Systems
    |January 24, 2017
    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

    Stress-Lensed Electrochemical Sintering Enables Fast and Stable Lithium-Silicon Alloy Chemistry in All-Solid-State Batteries.

    Advanced materials (Deerfield Beach, Fla.)·2026
    Same author

    LangSurf: Language-Embedded Surface Gaussians for 3D Scene Understanding.

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

    Breathing New Life into Small Object Detection with Detection-Oriented Rectification.

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

    PathTIGR: A pathway topology-informed graph representation learning framework for immunotherapy response prediction.

    Science advances·2026
    Same author

    Interpretable graph deep learning framework for drug synergy prediction by integrating functional and clinical similarities.

    NPJ digital medicine·2026
    Same author

    Ultrasound-Mediated Thrombolysis: From Mechanistic Insights to Advanced Nanoplatforms and Clinical Translation.

    Advanced healthcare materials·2026

    This study introduces mesh convolutional restricted Boltzmann machines (MCRBMs) and mesh convolutional deep belief networks (MCDBNs) for 3-D shape analysis. These novel deep learning models effectively learn structure-preserving features, outperforming existing methods in shape retrieval and correspondence tasks.

    Area of Science:

    • Computer Vision
    • Machine Learning
    • 3-D Shape Analysis

    Background:

    • Traditional 3-D feature learning methods struggle with human intervention, preserving local/global structures, irregular mesh topology, orientation ambiguity, and transformations.
    • Existing deep learning models are not directly applicable to irregular 3-D mesh data.

    Purpose of the Study:

    • To propose a novel deep learning model for learning discriminative 3-D mesh features.
    • To address the limitations of existing 3-D feature learning methods by preserving local and global structures and handling mesh irregularities.

    Main Methods:

    • Introduced mesh convolutional restricted Boltzmann machines (MCRBMs) with a novel irregular model structure.
    • Developed mesh convolutional deep belief networks (MCDBNs) by stacking MCRBMs.

    Related Experiment Videos

    Last Updated: Mar 8, 2026

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
    04:48

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    852
  • Employed a local structure preserving convolution (LSPC) strategy to handle orientation ambiguity and convolve learned structures.

    • Main Results:

    • MCRBMs learn structure-preserving local and global features from local function energy distribution.
    • MCDBNs effectively resolve orientation ambiguity on mesh surfaces.
    • Experimental results on global/partial shape retrieval and shape correspondence demonstrate superior performance compared to state-of-the-art methods.

    Conclusions:

    • The proposed MCRBM and MCDBN models offer a robust solution for 3-D shape analysis.
    • These methods significantly advance the state-of-the-art in learning discriminative features from 3-D meshes.