Authors:
J. Angelin Jeba, S. Rubin Bose, R. Regin, O. Jeba Singh, Bushra Rehman, R. Nathiya
Addresses:
Department of Electronics and Communication Engineering, S. A. Engineering College, Chennai, Tamil Nadu, India. School of Computer Science and Engineering, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Center for Academic Research, Alliance University, Bengaluru, Karnataka, India. Institute of Pathology and Diagnostic Medicine, Khyber Medical University, Hayatabad, Peshawar, Pakistan. Department of Computer Applications, Hindustan Institute of Technology and Science, Chennai, Tamil Nadu, India.
Abstract:
Utilising the potential of transfer learning in conjunction with deep convolutional neural networks, this research presents a comprehensive framework for the automated diagnosis of pneumonia based on chest radiographs. Therefore, the suggested methodology uses the VGG19 architecture to extract meaningful features from X-ray images, thereby improving diagnostic accuracy and generalisation. The model's robustness is further improved by extensive data augmentation and iterative fine-tuning, enabling consistent performance despite class imbalances and limited medical data. Experimental findings from a labelled chest X-ray dataset demonstrate the model's capacity to effectively differentiate between normal and pneumonia cases. These results indicate considerable improvements in classification accuracy, sensitivity, and specificity. This technique has the potential to assist clinical workflows, accelerate patient assessment, and promote timely action. It does this by streamlining detection operations and minimising the likelihood of human error. The expansion of algorithmic variety, optimisation of computing efficiency, and integration of multimodal data sources should be the focus of future research initiatives. This will pave the way for more scalable and intelligent diagnostic systems in the healthcare industry.
Keywords: Pneumonia Detection; Chest Radiographs; Transfer Learning; Healthcare Industry; Fine Tuning Strategies; Data Augmentation; Diagnostic Systems; Algorithmic Diversity.
Received: 19/12/2024, Revised: 07/04/2025, Accepted: 18/07/2025, Published: 09/12/2025
DOI: 10.64091/ATIHL.2025.000209
AVE Trends in Intelligent Health Letters, 2025 Vol. 2 No. 4 , Pages: 216-227