Related Experiment Video
Updated: Oct 4, 2025
04:17
DNA Virus Detection System Based on RPA-CRISPR/Cas12a-SPM and Deep Learning
Published on: May 10, 2024
937
An efficient hardware architecture based on an ensemble of deep learning models for COVID -19 prediction
Sakthivel R1, I Sumaiya Thaseen2, Vanitha M2
1School of Electronics Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
Sustainable Cities and Society
|February 9, 2022
Summary
This study introduces an efficient hardware architecture using an ensemble of deep learning models for COVID-19 detection from chest X-rays. The proposed system achieves high accuracy, improving COVID-19 diagnosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Engineering
Background:
- Deep learning models excel in image classification tasks.
- Current COVID-19 detection often relies on single deep learning models.
- Chest X-ray (CXR) analysis is crucial for diagnosing respiratory illnesses.
Purpose of the Study:
- To develop an efficient hardware architecture for COVID-19 identification using an ensemble deep learning model.
- To enhance the performance and accuracy of COVID-19 detection from CXR records.
- To optimize the hardware for real-time processing and reduced resource consumption.
Main Methods:
- Ensembling five deep learning models: ResNet, fitness, IRCNN, effectiveness, and Fitnet.
- Designing an application-specific hardware architecture with pipeline and parallel processing.
- Modeling the processing element (PE) and CNN architecture using Verilog and synthesizing with TSMC 90nm technology.
Main Results:
- Achieved high performance metrics for COVID-19 detection: 0.99 accuracy, 0.98 precision, 0.98 recall, and 0.98 F1-score.
- Demonstrated a 40% reduction in latency and clock cycles through hardware optimization.
- Minimized computations and power consumption by designing a data-aware PE.
Conclusions:
- The proposed ensemble deep learning hardware architecture significantly enhances COVID-19 detection accuracy from CXR.
- The optimized hardware design offers substantial improvements in processing speed and efficiency.
- This architecture is well-suited for rapid and reliable COVID-19 prediction and diagnosis.
