Search research articles

Search research articles

Related Experiment Video

Updated: Mar 7, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Deep-cascade: Cascading 3D Deep Neural Networks for Fast Anomaly Detection and Localization in Crowded Scenes.

Mohammad Sabokrou, Mohsen Fayyaz, Mahmood Fathy

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society

|February 22, 2017

Summary

This study introduces a fast method for detecting anomalies in crowded videos using a cascade of classifiers. The approach efficiently identifies unusual events, improving upon existing techniques in computation time.

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

Diffusion models in medical imaging: A comprehensive survey.

Medical image analysis·2023

Same author

Fast Weakly Supervised Action Segmentation Using Mutual Consistency.

IEEE transactions on pattern analysis and machine intelligence·2021

Same author

Deep End-to-End One-Class Classifier.

IEEE transactions on neural networks and learning systems·2020

Same author

Switching Structured Prediction for Simple and Complex Human Activity Recognition.

IEEE transactions on cybernetics·2020

Same author

A hierarchical, scalable architecture for a real-time monitoring system for an electrocardiography, using context-aware computing.

Journal of biomedical informatics·2019

Same author

Chemometrics and hyperspectral imaging applied to assessment of chemical, textural and structural characteristics of meat.

Meat science·2018

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

Area of Science:

Computer Science
Artificial Intelligence
Video Analytics

Background:

Anomaly detection in crowded scenes is computationally intensive.
Existing methods struggle with real-time processing for time-efficient localization.

Purpose of the Study:

To propose a novel, fast, and reliable method for anomaly detection and localization in crowded video data.
To address the challenge of time-efficient anomaly localization.

Main Methods:

A cubic-patch-based method utilizing a cascade of classifiers.
A two-stage approach: a 3D auto-encoder for initial normal patch identification, followed by a 3D Convolutional Neural Network (CNN) for deeper analysis.
Cascaded classifiers with shallow layers acting as Gaussian classifiers for simple normal patches and deeper layers for complex normal patches.

Related Experiment Videos

Last Updated: Mar 7, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Main Results:

The proposed method achieves performance comparable to top-performing anomaly detection and localization techniques on standard benchmarks.
Demonstrates superior performance in terms of required computation time compared to existing methods.
Effectively identifies and localizes anomalies in crowded video scenes.

Conclusions:

The novel cascade of two cascaded classifiers offers a time-efficient solution for anomaly detection and localization in crowded videos.
The method provides a reliable and fast alternative to current state-of-the-art approaches.
Advanced feature learning within the cascaded network architecture contributes to its effectiveness.