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 11, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Dual-Masked and Discriminative Reconstruction for Unified Vision Anomaly Detection.

Bin-Bin Gao

    IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
    |May 1, 2026
    PubMed
    Summary

    Dual-masked and Discriminative Reconstruction (D2Rec) tackles unified vision anomaly detection by preventing overfitting and enhancing discrimination. This simple, general method improves both anomaly classification and segmentation across diverse datasets.

    Related Concept Videos

    Masking and Demasking Agents01:19

    Masking and Demasking Agents

    4.1K
    EDTA titrations may necessitate masking and demasking agents to temporarily protect a particular metal ion in a mixture from the EDTA reaction. These agents facilitate the sequential analysis of the metal ions by forming stable complexes with some—but not all—metal ions during certain steps.
    There are many masking agents, such as cyanide, fluoride, triethanolamine, thiourea, and 2,3-bis(sulfanyl)propan-1-ol (formerly 2,3-dimercapto-1-propanol), with the masking agent chosen based on...
    4.1K

    You might also read

    Related Articles

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

    Sort by
    Same author

    Tumor microenvironment characteristics association with clinical outcome in patients with resected intestinal-type gastric cancer.

    The oncologist·2024
    Same author

    Increasing Sufu gene dosage reveals its unorthodox role in promoting polydactyly and medulloblastoma tumorigenesis.

    JCI insight·2024
    Same author

    Solid pseudopapillary neoplasm: Report of a case of primary ovarian origin and review of the literature.

    Heliyon·2023
    Same author

    Learning to Discover Multi-Class Attentional Regions for Multi-Label Image Recognition.

    IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2021
    Same author

    YTHDF2 is a potential target of AML1/ETO-HIF1α loop-mediated cell proliferation in t(8;21) AML.

    Oncogene·2021
    Same author

    Preparation of carbon-based AuAg alloy nanoparticles by using the heterometallic [Au<sub>4</sub>Ag<sub>4</sub>] cluster for efficient oxidative coupling of anilines.

    Dalton transactions (Cambridge, England : 2003)·2018

    Area of Science:

    • Computer Vision
    • Machine Learning
    • Artificial Intelligence

    Background:

    • Unsupervised reconstruction networks offer unified anomaly detection for image classification and segmentation.
    • Existing methods struggle with complex data distributions and overfitting due to "identity shortcut".
    • Current approaches often exhibit "weak discrimination", entangling normal and abnormal features, leading to inaccurate segmentation.

    Purpose of the Study:

    • To propose a simple yet general method, Dual-masked and Discriminative Reconstruction (D2Rec), for unified vision anomaly detection.
    • To address the challenges of "identity shortcut" and "weak discrimination" in reconstruction-based anomaly detection.
    • To enhance both anomaly classification and segmentation accuracy in an unsupervised manner.

    Main Methods:

    Related Experiment Videos

    Last Updated: Jun 11, 2026

    End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
    03:31

    End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

    Published on: December 15, 2023

    • Introduced a dual-masked reconstruction approach using complementary masks to prevent "identity shortcut".
    • Developed a self-supervised discriminator that refines reconstruction errors with synthesized anomalies to improve feature discrimination.
    • Designed D2Rec as a universal plugin compatible with various reconstruction-based anomaly detection architectures.

    Main Results:

    • D2Rec effectively resolves the "identity shortcut" by ensuring reconstruction relies on unmasked features.
    • The self-supervised discriminator significantly enhances the ability to distinguish between normal and abnormal features.
    • Outperformed previous methods on industrial (MVTec, BTAD, VisA) and medical (Brain MRI, Liver CT, Retinal OCT) datasets.

    Conclusions:

    • D2Rec provides a simple and effective solution for unified vision anomaly detection.
    • The proposed method demonstrates superior performance in both anomaly classification and segmentation.
    • D2Rec's universal plugin nature allows easy integration into existing reconstruction-based models.