Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Nov 26, 2025

Comparing Eye-tracking Data of Children with High-functioning ASD, Comorbid ADHD, and of a Control Watching Social Videos

Comparing Eye-tracking Data of Children with High-functioning ASD, Comorbid ADHD, and of a Control Watching Social Videos

Published on: December 7, 2018

Anomaly Detection With Bidirectional Consistency in Videos.

Zhiwen Fang, Jiafei Liang, Joey Tianyi Zhou

IEEE Transactions on Neural Networks and Learning Systems

|December 9, 2020

Summary

This study introduces a novel Siamese generative network (SIGnet) for anomaly detection. SIGnet uses mutual supervision to improve generalization and reduce false positives in normal frames, enhancing detection accuracy.

Related Concept Videos

Relative Motion Analysis using Rotating Axes

Relative Motion Analysis using Rotating Axes

Consider a component AB undergoing a linear motion. Along with a linear motion, point B also rotates around point A. To comprehend this complex movement, position vectors for both points A and B are established using a stationary reference frame.
However, to express the relative position of point B relative to point A, an additional frame of reference, denoted as x'y', is necessary. This additional frame not only translates but also rotates relative to the fixed frame, making it...

You might also read

Related Articles

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

Sort by

Same author

A comprehensive collection of systems biology data characterizing the host response to viral infection.

Scientific data·2015

Same author

7.6 W 1342 nm passively mode-locked picosecond composite Nd:YVO4/YVO4 laser with a semiconductor saturable absorber mirror.

Applied optics·2015

Same author

Millijoule-level picosecond mid-infrared optical parametric amplifier based on MgO-doped periodically poled lithium niobate.

Applied optics·2015

Same author

Structural basis of non-steroidal anti-inflammatory drug diclofenac binding to human serum albumin.

Chemical biology & drug design·2015

Same author

Involvement of cAMP-PKA pathway in adenosine A1 and A2A receptor-mediated regulation of acetaldehyde-induced activation of HSCs.

Biochimie·2015

Same author

Synergistic degradation of chitosan by impinging stream and jet cavitation.

Ultrasonics sonochemistry·2015

Area of Science:

Computer Science
Artificial Intelligence
Machine Learning

Background:

Anomaly detection often relies on self-supervised models trained with reconstruction constraints.
These models risk overfitting and sensitivity to normal patterns, leading to irregular responses.
Existing methods struggle with stable detection of anomalies in complex data.

Purpose of the Study:

To address limitations in current anomaly detection models, particularly overfitting and sensitivity to normal patterns.
To propose a novel mutual supervision framework for enhanced anomaly detection.
To improve the generalization ability and stability of anomaly detection systems.

Main Methods:

Developed a SIamese generative network (SIGnet) with two subnetworks for simultaneous forward and backward frame pattern modeling.

Related Experiment Videos

Last Updated: Nov 26, 2025

Comparing Eye-tracking Data of Children with High-functioning ASD, Comorbid ADHD, and of a Control Watching Social Videos

Comparing Eye-tracking Data of Children with High-functioning ASD, Comorbid ADHD, and of a Control Watching Social Videos

Published on: December 7, 2018

Implemented a bidirectional consistency loss as a regularization term to enhance model generalization.

Introduced a consistency-based evaluation criterion for stable scoring of normal frames.

Main Results:

The proposed SIGnet effectively models patterns in both forward and backward frames.
Bidirectional consistency loss significantly improved the model's generalization capabilities.
The consistency-based evaluation criterion ensured stable performance on normal frames, aiding anomaly identification.
Demonstrated superior effectiveness on several challenging benchmark datasets.

Conclusions:

The mutual supervision approach in SIGnet effectively mitigates overfitting and improves robustness.
SIGnet offers a stable and accurate solution for anomaly detection, outperforming existing methods.
The proposed method shows significant promise for real-world anomaly detection applications.