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: Dec 10, 2025

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
05:48

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception

Published on: August 9, 2024

1.9K

Revealing the Invisible with Model and Data Shrinking for Composite-database Micro-expression Recognition.

Zhaoqiang Xia, Wei Peng, Huai-Qian Khor

    IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
    |August 27, 2020
    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

    Ligand Geometry Regulated Architecture of Ultra-Microporous Flexible Guanidinium-Based Hydrogen-Bonded Organic Frameworks for Highly Selective Nitrous Oxide/Nitrogen Separation.

    Small (Weinheim an der Bergstrasse, Germany)·2026
    Same author

    Network-Texture-Induced Uniform Nucleation: Controllable Preparation and Application of High-Performance CsPbI<sub>3</sub> Nanocrystals in Al<sup>3+</sup>/Gd<sup>3+</sup> Co-Doped Glass.

    Inorganic chemistry·2026
    Same author

    Anomalous luminescence properties in Dy<sup>3+</sup>-doped Bi<sub>2</sub>O<sub>3</sub>-B<sub>2</sub>O<sub>3</sub>-SiO<sub>2</sub> glasses at high silver concentrations.

    Applied optics·2026
    Same author

    Is Coffee Consumption Associated With Increased Risk of Atrial Fibrillation: A Systematic Review and a Meta-Analysis.

    Pacing and clinical electrophysiology : PACE·2026
    Same author

    Design of phosphors in glass doped with silver nanocrystals (Sr,Ca)AlSiN<sub>3</sub>:Eu<sup>2+</sup> for sunlight-like lighting excited by violet light chips.

    Applied optics·2026
    Same author

    ResoPhys: Unsupervised Plug-and-Play Remote Physiological Measurement via Facial Videos of Arbitrary Resolution.

    IEEE journal of biomedical and health informatics·2026

    To improve micro-expression recognition using composite databases, researchers found that reducing input and model complexity helps deep models perform better. They developed a Recurrent Convolutional Network (RCN) with parameter-free modules to enhance performance on diverse datasets.

    Area of Science:

    • Computer Science
    • Artificial Intelligence
    • Machine Learning

    Background:

    • Composite databases enhance micro-expression recognition (MER) practicality but introduce domain shift challenges.
    • Deep models often suffer performance degradation due to domain shift in composite-database MER.
    • Learning complexity, including input and model complexity, significantly impacts deep model performance.

    Purpose of the Study:

    • To analyze the influence of learning complexity on deep model performance in composite-database MER.
    • To propose a novel Recurrent Convolutional Network (RCN) that mitigates performance degradation.
    • To introduce parameter-free modules that enhance representation ability without increasing model complexity.

    Main Methods:

    • Investigated the impact of lower-resolution input and shallower model architectures.

    More Related Videos

    Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm
    09:49

    Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm

    Published on: December 24, 2015

    14.5K

    Related Experiment Videos

    Last Updated: Dec 10, 2025

    Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
    05:48

    Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception

    Published on: August 9, 2024

    1.9K
    Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm
    09:49

    Holistic Facial Composite Creation and Subsequent Video Line-up Eyewitness Identification Paradigm

    Published on: December 24, 2015

    14.5K
  • Developed a Recurrent Convolutional Network (RCN) integrating shallower architectures and lower-resolution inputs.
  • Integrated three parameter-free modules (wide expansion, shortcut connection, attention unit) into the RCN.
  • Utilized a neural architecture search (NAS) strategy to combine modules for improved robustness.

    • Main Results:

    • Confirmed that lower-resolution input and shallower models ease performance degradation in composite-database tasks.
    • The proposed RCN with parameter-free modules demonstrated enhanced representation ability.
    • The NAS-optimized RCN architecture achieved superior robustness and performance.
    • Experiments on the MEGC2019 dataset showed state-of-the-art results compared to existing approaches.

    Conclusions:

    • Reducing learning complexity is crucial for effective deep learning in composite-database micro-expression recognition.
    • The proposed RCN framework with parameter-free modules offers a robust and efficient solution.
    • The integration of NAS further optimizes the architecture for enhanced performance and generalization.