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

Using functional magnetic resonance imaging to evaluate an acute allograft rejection model in rats.

Magnetic resonance imaging·2019
Same author

QsvR integrates into quorum sensing circuit to control Vibrio parahaemolyticus virulence.

Environmental microbiology·2019
Same author

mi R -15a/15b Cluster Modulates Survival of Mesenchymal Stem Cells to Improve Its Therapeutic Efficacy of Myocardial Infarction.

Journal of the American Heart Association·2019
Same author

Retrospective analysis of Clostridium difficile infection in patients with ulcerative colitis in a tertiary hospital in China.

BMC gastroenterology·2019
Same author

CCCH-type zinc finger antiviral protein is specifically overexpressed in spleen in response to subgroup J avian leukosis virus infection in chicken.

Research in veterinary science·2018
Same author

Mapping Twisted Light into and out of a Photonic Chip.

Physical review letters·2018

Related Experiment Video

Updated: Jun 27, 2025

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates
11:24

Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates

Published on: March 7, 2017

7.0K

ADPS: Asymmetric Distillation Postsegmentation for Image Anomaly Detection.

Peng Xing, Hao Tang, Jinhui Tang

    IEEE Transactions on Neural Networks and Learning Systems
    |April 29, 2024
    PubMed
    Summary

    This study introduces asymmetric distillation post-segmentation (ADPS) for anomaly detection, enhancing how student networks learn from teacher networks. ADPS significantly improves anomaly detection accuracy and segmentation, outperforming existing methods.

    More Related Videos

    Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration
    07:03

    Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration

    Published on: February 23, 2017

    7.7K
    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

    Related Experiment Videos

    Last Updated: Jun 27, 2025

    Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates
    11:24

    Evaporation-reducing Culture Condition Increases the Reproducibility of Multicellular Spheroid Formation in Microtiter Plates

    Published on: March 7, 2017

    7.0K
    Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration
    07:03

    Medical-grade Sterilizable Target for Fluid-immersed Fetoscope Optical Distortion Calibration

    Published on: February 23, 2017

    7.7K
    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

    Area of Science:

    • Computer Vision
    • Machine Learning
    • Artificial Intelligence

    Background:

    • Knowledge distillation-based anomaly detection (KDAD) uses teacher-student networks.
    • Existing KDAD methods have limitations in representation learning and feature utilization.

    Purpose of the Study:

    • To address limitations in current KDAD methods.
    • To propose an innovative asymmetric distillation post-segmentation (ADPS) approach.

    Main Methods:

    • ADPS utilizes an asymmetric distillation paradigm with distinct image forms for teacher-student networks.
    • A Weight Mask Block (WMB) generates a coarse anomaly localization mask.
    • The WMB transfers distilled knowledge back to the teacher network via a post-segmentation module (PSM).

    Main Results:

    • ADPS effectively detects and segments anomalies with fine structures and clear boundaries.
    • The method outperforms state-of-the-art anomaly detection techniques.
    • Significant improvements in average precision (AP) of 9% and 20% were observed on MVTec AD and KolektorSDD2 datasets, respectively.

    Conclusions:

    • The proposed ADPS method offers a novel and effective solution for anomaly detection and segmentation.
    • Asymmetric distillation and WMB integration enhance feature discrimination and knowledge transfer.
    • ADPS demonstrates superior performance compared to existing methods in benchmark evaluations.