Related Experiment Video
Updated: Jul 29, 2025
10:05
High-throughput Detection Method for Influenza Virus
Published on: February 4, 2012
26.3K
Flu-Net: two-stream deep heterogeneous network to detect flu like symptoms from videos using grey wolf optimization
Himanshu Gupta1, Javed Imran2, Chandani Sharma1
1Department of Computer Science and Engineering, Quantum University, Roorkee, India.
Journal of Ambient Intelligence and Humanized Computing
|May 25, 2023
Summary
This study introduces Flu-Net, an AI framework using CCTV footage to detect flu-like symptoms such as coughing and sneezing. The system achieved 70% accuracy, improving early detection and infection control.
Area of Science:
- Computer Science
- Artificial Intelligence
- Biomedical Engineering
Background:
- The global COVID-19 pandemic highlights the need for effective disease surveillance.
- Identifying flu-like symptoms early is crucial for limiting infectious disease spread.
Purpose of the Study:
- To develop an AI-powered framework, Flu-Net, for recognizing flu-like symptoms from surveillance video.
- To leverage human action recognition and deep learning for public health monitoring.
Main Methods:
- Utilized frame differencing to extract foreground motion from CCTV video.
- Employed a two-stream heterogeneous network (2D/3D ConvNets) for activity recognition.
- Integrated Grey Wolf Optimization (GWO) for feature selection.
Main Results:
- Flu-Net achieved 70% accuracy in identifying flu-like symptoms.
- The framework demonstrated an improvement of over 8% compared to baseline methods.
- Successfully recognized actions like coughing and sneezing in video data.
Conclusions:
- The proposed Flu-Net framework shows significant potential for real-time public health surveillance.
- AI-driven analysis of surveillance footage can enhance early detection of infectious disease outbreaks.
- This approach offers a scalable solution for monitoring public health in real-world settings.
