Year: 2026
Type: article
Abstract: Pulmonary diseases, including pneumonia, tuberculosis, asthma, COPD, fibrosis, emphysema, and COVID-19, pose significant diagnostic challenges.Traditional machine learning models often produce suboptimal classification accuracy.To tackle this challenge, an enhanced version of AlexNet, termed Improved AlexNet with Convolutional Neural Networks (IANET), is introduced as a deep transfer learning framework specifically optimized for pulmonary disease classification.IANET incorporates enhanced preprocessing techniques such as median and Wiener filtering and CLAHE enhancement, refining image quality for better diagnostic performance.Using NIH chest X-ray datasets and real-time hospital data, IANET achieved classification accuracies of 96% and 96.8%, respectively, surpassing existing methods.Unlike a direct application of AlexNet, IANET includes targeted architectural modifications-such as reduced kernel sizes, optimized stride values, and refined filter numbers-designed to preserve local image details while reducing complexity.Combined with our novel fusion filter, this customized approach significantly improves diagnostic accuracy for pulmonary disease classification.The results demonstrate IANET's potential in assisting specialists with precise, automated pulmonary disease diagnoses.
Authors: Ramadoss Ramalingam, Vimala Chinnaiyan
Institution: SRM Institute of Science and Technology
Language: English
Topic: COVID-19 diagnosis using AI
Subfield: Radiology, Nuclear Medicine and Imaging
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7147145192