PNEUMONIA DETECTION USING CHEST X-RAY
- Version
- Download 17
- File Size 241.18 KB
- File Count 1
- Create Date 27 February 2025
- Last Updated 27 February 2025
PNEUMONIA DETECTION USING CHEST X-RAY
Dr.P.Radha
Asst.Prof.,Department of Computer Science, Sri Krishna Arts and Science College,
Coimbatore. Email- radhap@skasc.ac.in
A.Bharanidharan
UG Student, Department of Computer Science, Sri Krishna Arts and Science College,
Coimbatore. Email – bharanidharansheeba@gmail.com
V.Madhumitha
UG Student, Department of Computer Science, Sri Krishna Arts and Science College,
Coimbatore. Email –madhuvelu1@gmail.com
ABSTRACT
Pneumonia continues to pose a significant challenge to global health, with its impact disproportionately affecting vulnerable populations, including young children, the elderly, and those with weakened immune systems. In many regions, there is a shortage of skilled radiologists, creating a critical need for faster and more reliable diagnostic tools that can help overcome these barriers. In response to this pressing issue, our study presents an innovative solution in the form of an automated pneumonia detection system, utilizing advanced machine learning (ML) techniques and chest X-ray images. By leveraging Convolutional Neural Networks (CNNs), a powerful form of deep learning, our system can detect even the most subtle signs of pneumonia in radiographs, which is often challenging for human eyes, particularly in the absence of expert radiologists.
CNNs are particularly suited for image analysis due to their ability to automatically learn complex hierarchical patterns from visual data. This means the model can effectively detect pneumonia without the need for manual feature extraction, making it more scalable and accurate. The model is trained using a comprehensive, labeled dataset of chest X-ray images, which enables it to identify and differentiate between healthy lung patterns and those affected by pneumonia. To rigorously assess the
performance of the system, we employ the F1 score, a metric that balances both precision and recall, ensuring the model maintains a high level of sensitivity in detecting true pneumonia cases while minimizing false positives. By using this balanced approach, we can confidently rely on the model to deliver accurate and actionable results in real-world medical environments.
The potential impact of this automated detection system extends far beyond improving diagnostic accuracy. By significantly reducing the time required to diagnose pneumonia, the system enables faster intervention and treatment, which is critical for improving patient outcomes. The consistency of results provided by the AI model also reduces the risk of human error, a concern in medical practice where incorrect or delayed diagnoses can have serious consequences. Moreover, this system has the potential to support healthcare professionals in resource-constrained settings, where access to skilled radiologists may be limited, providing a much-needed tool to enhance decision-making and improve patient care. As healthcare systems worldwide increasingly turn to AI-driven solutions, our research contributes to the ongoing transformation of the industry, paving the way for more efficient, accessible, and equitable healthcare that can tackle global health challenges like pneumonia with greater speed and accuracy.
Keywords – Pneumonia, Streptococcus pneumonia, Convolutional Neural Network, respiratory infections, X-ray images
Download