Anomaly Detection in Chest X-ray Images and Classifying Disease using  Deep Learning models

Chevella Aravind Reddy; Kulpaguri Rahul; Yelemela Shashank

doi:10.71097/IJSAT.v16.i2.5108

Anomaly Detection in Chest X-ray Images and Classifying Disease using Deep Learning models

Author(s)	Chevella Aravind Reddy, Kulpaguri Rahul, Yelemela Shashank
Country	India
Abstract	In this research we have explored effectiveness of autoencoder in anomalous detection, deep learning models in image classification and Grad-CAM in Explanation of classification and detection. We mainly focused on classifying Tuberculosis and Pneumonia. We collected normal, Tuberculosis and Pneumonia chest X-ray images from publicly available Kaggle dataset. Our research constitutes of two stages. In stage 1, We trained Autoencoder on only one class i.e., normal chest X-ray images. Thus, making the model more effective in detecting anomaly in a chest X-ray image if it is having anomalous pattern compared to normal image. If the model detects any anomaly it is directed to stage 2 where the deep learning models i.e., InceptionV3 and Xception are trained on Tuberculosis and Pneumonia chest X-ray images. These models are integrated with Grad-CAM to generate heatmap that detects and visualizes disease spot in the test image. These models are evaluated on training and testing accuracies. Xception model has achieved an accuracy of 100 % while InceptionV3 achieved an accuracy of 99.77%. The Integration of Grad-CAM made it easier to interpret and trust the model’s decision-making process.
Keywords	InceptionV3, Xception, Autoencoder, Grad-CAM
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 16, Issue 2, April-June 2025
Published On	2025-05-21
DOI	https://doi.org/10.71097/IJSAT.v16.i2.5108
Short DOI	https://doi.org/g9mn99

View / Download PDF File

About IJSAT Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	Message on WhatsApp	+91-9687-182-185	editor@ijsat.org

International Journal on Science and Technology

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Anomaly Detection in Chest X-ray Images and Classifying Disease using Deep Learning models

Share this