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: Apr 2, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

1.3K

LAE-Net: Large Pretrained Models Assistant Text-Guided Image Editing Adversarial Network.

Bing Yang, Xueqin Xiang, Yong Peng

    IEEE Transactions on Visualization and Computer Graphics
    |March 31, 2026
    PubMed
    Summary

    This study introduces LAE-Net, a novel approach for efficient text-guided image editing. It overcomes limitations of diffusion models by enabling faster, more integrated edits without redundant token learning.

    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

    PK-PD Integration Modeling and Cutoff Value of Florfenicol against <i>Streptococcus suis</i> in Pigs.

    Frontiers in pharmacology·2018
    Same author

    Dislodgement of circular mapping catheter electrode in the left atrium: A near miss.

    HeartRhythm case reports·2018
    Same author

    A Switchable Helical Capsule for Encapsulation and Release of Potassium Ion.

    The Journal of organic chemistry·2018
    Same author

    Infection Function of Adhesin-Like Protein ALP609 from Spiroplasma melliferum CH-1.

    Current microbiology·2018
    Same author

    Tumor-derived exosomes antagonize innate antiviral immunity.

    Nature immunology·2018
    Same author

    Preparation of Cyano-Substituted Tetraphenylethylene Derivatives and Their Applications in Solution-Processable OLEDs.

    Molecules (Basel, Switzerland)·2018

    Area of Science:

    • Computer Vision
    • Artificial Intelligence
    • Natural Language Processing

    Background:

    • Diffusion models excel at realistic image editing but are slow due to iterative evaluations.
    • Existing methods require unique textual tokens per image, leading to model fragmentation.

    Purpose of the Study:

    • To propose LAE-Net (Large pretrained models Assistant text-guided image Editing adversarial Network) for efficient text-guided image editing.
    • To address the speed and fragmentation issues of current image editing models.

    Main Methods:

    • Introduced a deep semantic editing network for global text information transfer.
    • Developed a discriminator using CLIP and DINO visual embeddings for enhanced discrimination.
    • Enabled differentiation of text-relevant and irrelevant image areas.

    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

    3.4K

    Related Experiment Videos

    Last Updated: Apr 2, 2026

    Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
    03:14

    Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

    Published on: December 6, 2024

    1.3K
    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

    3.4K

    Main Results:

    • LAE-Net demonstrates outstanding performance in text-guided image editing.
    • The proposed model surpasses several state-of-the-art image editing methods.
    • Achieved efficient and integrated image editing capabilities.

    Conclusions:

    • LAE-Net offers a significant advancement in automatic real image editing.
    • The model provides a more efficient and robust solution compared to existing approaches.
    • LAE-Net facilitates high-quality image manipulation through natural language guidance.