• Version
• Download 13
• File Size 424.35 KB
• File Count 1
• Create Date 20 November 2025
• Last Updated 20 November 2025

Detection Of Chest Diseases Using EfficientNet-B0

 

 

Varun V1, Adithya M2, Nagatejas P3, Nethra Moorthi Hosahalli4, Emmanuel Didymus Sebastian5 & Dr. Gowthul Alam6

1 2 3 4 5 6Department of Computer Science and Engineering, JAIN (Deemed-to-be-University)

 

 

Abstract - Chest X-rays are vital diagnostic tools in identifying thoracic diseases, but interpreting them manually can be time-consuming and prone to variability. With the advancement of deep learning, automated multi-label classification has become increasingly feasible for aiding clinical diagnosis. In this research, we investigate the performance of EfficientNetB0, a lightweight convolutional neural network, for classifying 14 chest diseases using the ChestMNIST dataset. The dataset comprises grayscale X-ray images, and the model is trained from scratch without using pre-trained weights to better adapt to the grayscale nature of the input. This research showcase that that EfficientNetB0 can effectively handle multi-label classification in chest radiography, even without leveraging transfer learning, making it a promising candidate for real-world clinical applications. The model provides a high accuracy and promising Area Under the Curve (AUC) value.

 

 

Key Words:  ChestMNIST, AUC, EfficientNetB0, Chest Disease, Transfer Learning

Download