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: Jan 15, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

1.0K

Interpretable facial landmark detection by multi-expert collaborative uncertainty-aware deep networks.

Jun Wan1, Hui Xi2, Yuanzhi Yao2

  • 1School of Information Engineering, Zhongnan University of Economics and Law, Wuhan, China; College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, China.

Neural Networks : the Official Journal of the International Neural Network Society
|October 16, 2025
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

Multiscale Organization of Neural Networks in a 3D Bioprinted Matrix.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2025
Same author

Spatial distribution patterns and formation of global spermatophytes.

Journal of integrative plant biology·2025
Same author

P-Band Center Engineering of Sulfur to Weaken Fe─S Bond Toward Superior Sodium-Ion Storage.

Advanced materials (Deerfield Beach, Fla.)·2025
Same author

Dark nectar pouches are visually similar to colored nectar in bird-pollinated flowers.

Ecology·2025
Same author

An in vivo study of the ameliorative effect of supplementation with Lacticaseibacillus paracasei Glory LP16 in immunocompromised mice.

Scientific reports·2025
Same author

Elevational variation in trichomes of the alpine subnival woolly plant Eriophyton wallichii: abiotic and biotic correlates and impacts on other traits.

Annals of botany·2025

This study introduces a novel Multi-Expert Collaborative Uncertainty-Aware Deep Network (MCUDN) for facial landmark detection (FLD). The MCUDN improves accuracy and interpretability by considering landmark uncertainty and enhancing facial shape constraints.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Artificial Intelligence

Background:

  • Heatmap regression methods dominate facial landmark detection (FLD).
  • Current methods optimize Gaussian distributions, neglecting landmark uncertainty.
  • Existing approaches struggle with large poses and occlusions due to weak facial shape constraints.

Purpose of the Study:

  • To develop a robust and interpretable deep network for facial landmark detection.
  • To address limitations of existing heatmap regression methods in handling uncertainty and complex facial variations.

Main Methods:

  • Proposed a Multi-Expert Collaborative Uncertainty-Aware Deep Network (MCUDN).
  • Introduced Uncertainty-Aware Regression (UAR) to adaptively weight landmarks based on uncertainty.
Keywords:
Face alignmentFacial shape constraintsGlobal dependencyMulti-expertUncertainty-aware

More Related Videos

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K

Related Experiment Videos

Last Updated: Jan 15, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

1.0K
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K
  • Developed Multi-Expert Collaborative Learning (MECL) for enhanced facial shape modeling.

    • Main Results:

    • The UAR method dynamically controls localization gradients by penalizing uncertain landmarks.
    • MECL extracts collaborative features, strengthening facial shape constraints.
    • The integrated MCUDN framework demonstrated superior performance on benchmark datasets.

    Conclusions:

    • The MCUDN framework, combining UAR and MECL, significantly enhances facial landmark detection robustness and interpretability.
    • The proposed methods outperform current state-of-the-art approaches, particularly under challenging conditions.
    • This work offers a more accurate and interpretable solution for facial landmark detection.