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: May 14, 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

An explainable deep learning framework for video violence detection using unsupervised keyframe selection and

Rashid Azim¹, Naveed Abbas¹, Hend Khalid Alkahtani²

¹Department of Computer Science, Islamia College Peshawar, Peshawar, 25100, Pakistan.

Scientific Reports

|February 26, 2026

Summary

Related Concept Videos

Force Classification

Force Classification

Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...

You might also read

Related Articles

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

Sort by

Same author

A framework based on hybrid Suzuki-Abe and convex Hull approach for improved classification of skin lesions.

Scientific reports·2026

Same author

Multiobjective framework for hardware and quality aware approximate Gaussian filtering toward energy efficient ultrasound image denoising.

Scientific reports·2026

Same author

Detection and mitigation of abusive web traffic using convolutional neural networks.

Scientific reports·2026

Same author

Artificial intelligence with deep learning driven entropy-curvature attention mechanism for detection and segmentation of skin lesions using biomedical images.

Scientific reports·2026

Same author

Integrating dual convolutional networks and BiLSTM for precision prediction of chronic myeloid leukemia from protein sequences.

Frontiers in genetics·2026

Same author

An embedded deep learning framework for real-time violence detection and alert generation.

Scientific reports·2026

This summary is machine-generated.

This study introduces an Explainable Attention-Enhanced Convolutional Neural Network (CNN) for real-time video violence detection. The novel framework achieves high accuracy and efficiency, offering a transparent solution for surveillance systems.

Area of Science:

Computer Vision
Artificial Intelligence
Machine Learning

Background:

Increasing video data necessitates intelligent, explainable systems for real-time violence detection.
Existing methods face challenges with redundancy, transparency, and generalization.

Purpose of the Study:

To propose a novel Explainable Attention-Enhanced Convolutional Neural Network (CNN) framework for video violence detection.
To address redundancy, transparency, and generalization issues in automated violence detection.

Main Methods:

Unsupervised keyframe selection using similarity-based clustering to reduce computational load.
Attention modules integrated into the CNN for enhanced spatial-temporal feature learning.
Grad-CAM++ for interpretable visual insights into model decisions.

Keywords:

Attention-enhanced CNN Explainable deep learning Grad-CAM + + visualization Keyframe selection Violence detection

Related Experiment Videos

Last Updated: May 14, 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:

Achieved superior performance with an average accuracy of 94.6% and F1-score of 93.9% on five benchmark datasets.
Outperformed state-of-the-art models including C3D, I3D, ResNet-LSTM, and ViViT.
Demonstrated near-real-time efficiency (≈62 FPS) and reduced memory usage (6.8 GB).

Conclusions:

The proposed framework offers a robust, efficient, and transparent solution for automated video violence detection.
Keyframe selection and attention mechanisms significantly improve model performance.
Interpretability enhances reliability by highlighting violence-relevant regions, suitable for surveillance and public safety.