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: Nov 25, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
09:47

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

1.5K

MAMA Net: Multi-Scale Attention Memory Autoencoder Network for Anomaly Detection.

Yurong Chen, Hui Zhang, Yaonan Wang

    IEEE Transactions on Medical Imaging
    |December 16, 2020
    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

    mTORC1 signaling requires proteasomal function and the involvement of CUL4-DDB1 ubiquitin E3 ligase.

    Cell cycle (Georgetown, Tex.)·2008
    Same author

    Prospective study of liver transplant recipients with HCV infection: evidence for a causal relationship between HCV and insulin resistance.

    Liver transplantation : official publication of the American Association for the Study of Liver Diseases and the International Liver Transplantation Society·2008
    Same author

    Quantitative gel electrophoresis: sources of variation.

    Journal of proteome research·2008
    Same author

    Evidence that the Nijmegen breakage syndrome protein, an early sensor of double-strand DNA breaks (DSB), is involved in HIV-1 post-integration repair by recruiting the ataxia telangiectasia-mutated kinase in a process similar to, but distinct from, cellular DSB repair.

    Virology journal·2008
    Same author

    [Inhibitory effects of Qushi Huayu Decoction on fatty deposition and tumor necrosis factor alpha secretion in HepG2 cells induced by free fatty acid].

    Zhongguo Zhong xi yi jie he za zhi Zhongguo Zhongxiyi jiehe zazhi = Chinese journal of integrated traditional and Western medicine·2008
    Same author

    Bioactive polybrominated diphenyl ethers from the marine sponge Dysidea sp.

    Journal of natural products·2008
    Same journal

    MUST: Multi-style virtual staining with incomplete pairs.

    IEEE transactions on medical imaging·2026
    Same journal

    BrainCL: Transformer-Based Brain Network Contrastive Learning with Multi-Order Topology and Salience Masking.

    IEEE transactions on medical imaging·2026
    Same journal

    LLM-enhanced Neuron Segmentation and Reconstruction in Complex Mouse Brain Images.

    IEEE transactions on medical imaging·2026
    Same journal

    Matrixed-Spectrum Decomposition Accelerated Linear Boltzmann Transport Equation Solver for Fast Scatter Correction in Multi-Spectral CT.

    IEEE transactions on medical imaging·2026
    Same journal

    The Ritz Adjoint Method for MRI Pulse Design.

    IEEE transactions on medical imaging·2026
    Same journal

    Physiology-guided Self-supervised Learning for Simultaneous Dual-Tracer PET Separation.

    IEEE transactions on medical imaging·2026
    See all related articles

    This study introduces MAMA Net, a novel anomaly detection method using multi-scale attention and memory. It effectively identifies outliers by improving reconstruction accuracy and generalization across diverse datasets.

    Area of Science:

    • Computer Science
    • Artificial Intelligence
    • Medical Imaging Analysis

    Background:

    • Anomaly detection is crucial across domains but faces challenges with diverse outliers and opaque inference.
    • Existing methods like object detection and reconstruction error techniques have limitations in handling real-world complexities.

    Purpose of the Study:

    • To propose a novel Multi-scale Attention Memory with hash addressing Autoencoder network (MAMA Net) for robust anomaly detection.
    • To address limitations of existing methods by incorporating multi-scale spatial attention and a hash-based memory module.

    Main Methods:

    • Developed a multi-scale global spatial attention block to overcome convolution operator limitations and enhance feature representation.
    • Designed a hash addressing memory module to ensure anomalies yield higher reconstruction errors, improving classification.

    More Related Videos

    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

    817

    Related Experiment Videos

    Last Updated: Nov 25, 2025

    Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
    09:47

    Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

    Published on: December 15, 2023

    1.5K
    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

    817
  • Integrated Mean Square Error (MSE) with Wasserstein loss to refine encoding data distribution.

    • Main Results:

    • The MAMA Net demonstrated robust performance and excellent generalization capabilities.
    • Experiments on COVID-19 and brain MRI datasets validated the network's effectiveness.
    • The multi-scale attention block achieved competitive results with minimal network levels.

    Conclusions:

    • MAMA Net offers a significant advancement in anomaly detection by effectively handling diverse outliers.
    • The proposed architecture provides a powerful and generalizable solution for identifying anomalies in complex datasets.
    • The integration of attention mechanisms and memory modules enhances the interpretability and accuracy of anomaly detection.