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 Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Multi-Directional Long-Term Recurrent Convolutional Network for Road Situation Recognition.

Sensors (Basel, Switzerland)·2024
Same author

Lactate promotes fatty acid oxidation by the tricarboxylic acid cycle and mitochondrial respiration in muscles of obese mice.

American journal of physiology. Cell physiology·2024
Same author

Lupenone attenuates thapsigargin-induced endoplasmic reticulum stress and apoptosis in pancreatic beta cells possibly through inhibition of protein tyrosine kinase 2 activity.

Life sciences·2023
Same author

Whole body mass estimates and error propagation in countermovement jump: a simulated error study.

Sports biomechanics·2023
Same author

Enhancing Quality Control in Web-based Participatory Augmented Reality Business Card Information System Design.

Sensors (Basel, Switzerland)·2023
Same author

Early Life Stress Detection Using Physiological Signals and Machine Learning Pipelines.

Biology·2023

Related Experiment Video

Updated: Nov 18, 2025

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

3.1K

Portrait Segmentation Using Ensemble of Heterogeneous Deep-Learning Models.

Yong-Woon Kim1, Yung-Cheol Byun2, Addapalli V N Krishna3

  • 1Centre for Digital Innovation, CHRIST (Deemed to be University), Bangalore, Karnataka 560029, India.

Entropy (Basel, Switzerland)
|February 10, 2021
PubMed
Summary

Ensemble methods combining multiple deep learning models enhance portrait segmentation accuracy. This approach improves performance over single models, offering a more efficient solution for various applications.

Keywords:
deep learningefficiencyensembleportrait segmentationsimple soft votingstackingweighted soft voting

More Related Videos

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.9K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

10.0K

Related Experiment Videos

Last Updated: Nov 18, 2025

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

3.1K
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.9K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

10.0K

Area of Science:

  • Computer Vision
  • Artificial Intelligence
  • Machine Learning

Background:

  • Image segmentation is crucial for applications like medical imaging and autonomous vehicles.
  • Portrait segmentation, a key part of semantic image segmentation, is vital for security and video conferencing.
  • Deep learning has significantly advanced image segmentation, but existing models are often single-component.

Purpose of the Study:

  • To develop and evaluate an ensemble method combining heterogeneous deep learning models for improved portrait segmentation.
  • To assess the performance and cost-efficiency of ensemble models compared to single models.

Main Methods:

  • Proposed an ensemble approach integrating multiple deep learning portrait segmentation models.
  • Experimented with Two-Models and Three-Models ensembles using soft voting and weighted soft voting.
  • Evaluated performance using Intersection over Union (IoU), IoU standard deviation, and false prediction rate.

Main Results:

  • Ensemble models demonstrated higher accuracy and lower error rates than single deep learning models.
  • Ensemble methods generally increased memory and computing power requirements.
  • Specific ensemble configurations showed improved accuracy with reduced resource usage compared to single models.

Conclusions:

  • Ensemble deep learning models offer a promising strategy for enhancing portrait segmentation performance.
  • Balancing accuracy and computational cost is key when designing ensemble segmentation systems.
  • The proposed ensemble method provides a more robust and accurate solution for portrait segmentation tasks.