International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 12 | Dec 2025
p-ISSN: 2395-0072
www.irjet.net
Deep Learning-Based Automated Detection of Pneumonia from Chest X-Ray Using Convolutional Neural Network Md. Mayn Uddin Dept. of Electrical and Electronic Engineering, Jatiya Kabi Kazi Nazrul Islam University, Trishal, Mymensingh2224 , Bangladesh ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Pneumonia is considered a leading cause of childhood mortality worldwide, particularly in resource-limited settings where early diagnosis is challenging. Chest X-ray (CXR) imaging is the most widely used diagnostic tool, but its interpretation is highly prone to variability and error. To demonstrate this, we introduced a computer-aided diagnosis (CAD) system that leverages convolutional neural networks (CNNs) with batch normalization and transfer learning for automated pneumonia detection. Using the publicly available Kermany dataset, we pre-processed and resized CXR images, applied model checkpoints to minimize over fitting, and evaluated performance with a confusion matrix. Our proposed methodology achieved an accuracy of 85%, exhibiting that deep learning can provide reliable, scalable support for radiologists in distinguishing pneumonic from healthy lungs with early response. This work underscores the potential of CNN-based CAD systems to improve diagnostic consistency and accessibility in clinical practice. Key Words: Convolutional Neural Network, confusion matrix, keras, recall, hyper parameters, Pneumonia detection.
1. INTRODUCTION Pneumonia is a serious infectious disease of the respiratory system that arises from bacterial, viral, or fungal pathogens. The infection primarily affects the lungs by inducing inflammation in the alveoli and may lead to the accumulation of fluid, a condition commonly referred to as pleural effusion. Globally, pneumonia remains one of the leading causes of mortality among children below five years of age, contributing to more than fifteen percent of deaths in this age group [1]. The burden of the disease is particularly severe in low- and middle-income countries, where factors such as dense population, environmental pollution, poor hygiene, and limited access to healthcare facilities significantly increase vulnerability. Consequently, timely identification and appropriate treatment are essential to reduce the risk of fatal outcomes. Medical imaging techniques, including computed tomography (CT), magnetic resonance imaging (MRI), and chest radiography, are routinely employed to assist in the diagnosis of pneumonia. Among these modalities, chest X-ray imaging is widely preferred due to its non-invasive nature, affordability, and accessibility. Figure 1 illustrates representative chest X-ray images of healthy and pneumonic lungs. Infected lungs typically exhibit abnormal opacities, known as infiltrates, which appear as white regions and serve as key indicators of pneumonia. Despite their clinical usefulness, chest X-ray interpretations are often influenced by the experience and subjective judgment of radiologists, which can lead to diagnostic inconsistencies [2]. This limitation highlights the need for reliable automated approaches to support pneumonia detection. To address this challenge, the present study proposes a computer-aided diagnosis (CAD) framework based on an ensemble of deep transfer learning models for accurate classification of chest X-ray images. Such systems have the potential to assist clinicians by providing consistent and objective diagnostic support. Deep learning has emerged as a powerful branch of artificial intelligence, demonstrating remarkable performance in a wide range of computer vision applications [3]. In particular, convolutional neural networks (CNNs) have achieved significant success in image classification tasks across multiple domains [4]. However, the effectiveness of CNN-based models is highly dependent on the availability of large, well-annotated datasets. In the biomedical domain, obtaining extensive labeled image collections is challenging, as expert annotation by medical professionals is both costly and time-consuming. Transfer learning offers a practical solution to this limitation by enabling the reuse of knowledge from models trained on large-scale datasets. In this approach, pre-trained CNNs, commonly trained on datasets such as Image Net, which contains millions of natural images, are adapted to medical imaging tasks, thereby improving performance even when only limited training data are available.
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 1263