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: Jun 29, 2025

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

530

Transformer Based Pluralistic Image Completion With Reduced Information Loss.

Qiankun Liu, Yuqi Jiang, Zhentao Tan

    IEEE Transactions on Pattern Analysis and Machine Intelligence
    |April 2, 2024
    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

    Quercetin attenuates skin inflammation and fibrosis in systemic sclerosis by targeting the RELA/c-Jun axis to suppress th17 cell responses.

    Frontiers in immunology·2026
    Same author

    Interface engineering constructs Co-O<sub>V</sub>-Ce/La interfacial sites with dual "capture-clear" functionality to enhance water-resistant CO oxidation performance of Co<sub>3</sub>O<sub>4</sub> catalysts.

    Journal of colloid and interface science·2026
    Same author

    Enhancing Generative Models for Modality Imputation of 3-D MRIs via Consistency-Aware Refinement and Super-Resolution Guidance.

    IEEE transactions on neural networks and learning systems·2026
    Same author

    A 2-GS/s 35.9-fJ/conv.-step Voltage-Time Hybrid Pipelined ADC with Digital Background Calibration in 28-nm CMOS.

    Micromachines·2026
    Same author

    A prediction model for the risk of developing HFpEF during hospitalization in patients with acute myocardial infarction.

    International journal of cardiology. Heart & vasculature·2026
    Same author

    Synergistic effects of plaque geometry and composition on coronary hemodynamics and mechanical stability: a multiscale computational study.

    Biomedical physics & engineering express·2026
    Same journal

    TraGraph-GS: Trajectory Graph-based Gaussian Splatting for Arbitrary Large-Scale Scene Rendering.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    SWIFT: A Small-World Interaction Framework for Flow-Aware Trajectory Prediction in Autonomous Driving.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    HardFlow: Hard-Constrained Sampling for Flow-Matching Models Via Trajectory Optimization.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    Industrial Brain: Self-Evolving Neuro-Symbolic Autonomy with Causal Resilience for Cyber-Physical Systems.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    Adaptive Hardness-Driven Dictionary Distillation for Incomplete Streaming View Clustering.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same journal

    Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation.

    IEEE transactions on pattern analysis and machine intelligence·2026
    See all related articles

    This study introduces PUT, a novel transformer framework for image inpainting. It addresses information loss by using patch tokens and unquantized features, significantly improving image fidelity and diversity.

    Area of Science:

    • Computer Vision
    • Artificial Intelligence

    Background:

    • Transformer-based methods dominate image inpainting.
    • Existing methods suffer information loss due to downsampling and pixel value quantization.

    Purpose of the Study:

    • To propose a new transformer framework, PUT, to mitigate information loss in image inpainting.
    • To enhance image fidelity, diversity, and controllability in image restoration tasks.

    Main Methods:

    • Introduced a patch-based auto-encoder (P-VQVAE) to convert images into patch tokens, avoiding downsampling.
    • Employed an Un-quantized Transformer that processes P-VQVAE encoder features directly, bypassing quantization.
    • Incorporated semantic and structural conditions for guided inpainting.

    Main Results:

    More Related Videos

    A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
    04:23

    A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

    Published on: April 21, 2023

    1.8K
    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
    04:48

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    395

    Related Experiment Videos

    Last Updated: Jun 29, 2025

    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

    530
    A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
    04:23

    A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

    Published on: April 21, 2023

    1.8K
    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
    04:48

    Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

    Published on: July 5, 2024

    395
    • PUT significantly outperforms existing transformer-based methods in image fidelity.
    • Achieved superior diversity and fidelity compared to state-of-the-art pluralistic inpainting methods on large-scale datasets like ImageNet.

    Conclusions:

    • The proposed PUT framework effectively addresses information loss in transformer-based image inpainting.
    • PUT offers a more robust and controllable solution for high-fidelity image restoration and generation.