Interpretability of Convolutional Neural Network-based Models in Medical Image Analysis: A Study on RSNA Pneumonia Dataset for Pneumonia Detection.

In the realm of medical image analysis, machine learning (ML) holds significant promise for enhancing disease diagnosis, particularly with chest X-rays. However, the 'black box' nature of many ML models often generates skepticism among medical professionals, challenging the integration of these tools into clinical practice. This research addresses this issue by focusing on the interpretability of ML models used in pneumonia diagnosis. Utilizing the RSNA Pneumonia Dataset, the study explores advanced interpretability methods to provide clearer insights into model decisions without compromising accuracy. Through the application of contemporary techniques, the research aims to achieve a harmonious balance between model interpretability and diagnostic precision, ultimately contributing to more transparent and effective pneumonia diagnosis tools. 

Team members: MD ABRAR HASNAT, Md Jobayer, Md. Mehedi Hasan Shawon, Sumaiya Akter.