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

2s-DAS: Two-Stream Diffusion with Multi-Modal Fusion for Temporal Action Segmentation.

Journal of imaging·2026
Same author

Sliding physical invariant neural operator for long-term prediction of complex dynamics in physical systems.

National science review·2026
Same author

Integrating machine learning-based molecular design with experimental validation for the discovery of EGFR inhibitors in lung cancer.

Molecular diversity·2026
Same author

Ferroelectric-Polarization-Modulated 2D Floating-Gate Memory Enabling a 10<sup>6</sup> On/Off Ratio under ±1 V Gate-Voltage Sweep.

Nano letters·2026
Same author

Enriched Environment Suppresses Neuronal Ferroptosis Through SIRT1/AKT/GSK3β-Dependent Glycogen Metabolic Reprogramming After Cerebral Ischemia-Reperfusion.

Antioxidants (Basel, Switzerland)·2026
Same author

MS-PANet: Multi-Scale Spatial Pyramid Attention for Effective Drainage Pipeline Image Dehazing.

Journal of imaging·2026

Related Experiment Video

Updated: Jun 10, 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

2.7K

Domain adaptive semantic segmentation by optimal transport.

Yaqian Guo1, Xin Wang1, Ce Li2

  • 1Department of Mathematics, Shanghai University, Shanghai 200444, China.

Fundamental Research
|October 21, 2024
PubMed
Summary
This summary is machine-generated.

This study introduces a domain adaptation framework using optimal transport (OT) and attention mechanisms for semantic scene segmentation with limited labeled data. The method significantly improves mean intersection-over-union (mIOU) performance in autonomous driving scenarios.

Keywords:
Deep learningMultiscale networkOptimal transportSemantic scene segmentationUnsupervised domain adaptation

More Related Videos

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

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

370
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

8.9K

Related Experiment Videos

Last Updated: Jun 10, 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

2.7K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

370
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
08:25

Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment

Published on: May 7, 2019

8.9K

Area of Science:

  • Computer Vision
  • Machine Learning
  • Autonomous Driving

Background:

  • Semantic scene segmentation is crucial for autonomous driving perception, assigning labels to image pixels.
  • Current methods, often CNN-based, require extensive labeled data, limiting their practical application.
  • Developing effective segmentation with minimal labeled data is a key research challenge.

Purpose of the Study:

  • To develop a domain adaptation framework for semantic scene segmentation using limited labeled data.
  • To leverage optimal transport (OT) and attention mechanisms for robust domain alignment.
  • To enhance feature representation and multiscale properties for improved segmentation accuracy.

Main Methods:

  • A CNN generates the output space for feature representation.
  • Optimal transport (OT) aligns source and target domains in the output space, guided by an attention mechanism.
  • A multiscale segmentation network is employed to capture diverse feature properties.

Main Results:

  • The proposed OT-based domain adaptation framework significantly improved mean intersection-over-union (mIOU).
  • The method demonstrated superior performance compared to benchmark and state-of-the-art (SOTA) semantic segmentation techniques.
  • Visualization results confirmed enhanced performance across various domain adaptation scenarios.

Conclusions:

  • The developed framework effectively addresses the challenge of semantic scene segmentation with limited labeled data.
  • Optimal transport and attention mechanisms provide a robust and interpretable approach to domain adaptation.
  • The multiscale segmentation network enhances the model's ability to handle complex scene properties.